36305 Disease and Mortality in Sub-Saharan Africa Second Edition Edited by Dean T. Jamison, Richard G. Feachem, Malegapuru W. Makgoba, Eduard R. Bos, Florence K. Baingana, Karen J. Hofman, and Khama O. Rogo Disease and Mortality in Sub-Saharan Africa Second Edition Disease and Mortality in Sub-Saharan Africa Second Edition Editors Dean T. Jamison Richard G. Feachem Malegapuru W. Makgoba Eduard R. Bos Florence K. Baingana Karen J. Hofman Khama O. Rogo THE WORLD BANK Washington, D.C. ©2006 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved 1 2 3 4 09 08 07 06 This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work with- out permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete infor- mation to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street, NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN-10: 0-8213-6397-2 ISBN-13: 978-0-8213-6397-3 eISBN: 0-8213-6398-0 DOI: 10.1596/978-0-8213-6397-3 Library of Congress Cataloging-in-Publication data has been applied for. Contents Foreword xiii Acknowledgments xv Contributors xvii Abbreviations and Acronyms xxi Chapter 1 Changing Patterns of Disease and Mortality in Sub-Saharan Africa: An Overview 1 Florence K. Baingana and Eduard R. Bos Chapter 2 Levels and Trends in Mortality in Sub-Saharan Africa: An Overview 11 Jacob Adetunji and Eduard R. Bos Chapter 3 Trends in Child Mortality, 1960 to 2000 15 Kenneth Hill and Agbessi Amouzou Chapter 4 Levels and Trends of Adult Mortality 31 Debbie Bradshaw and Ian M. Timaeus Chapter 5 Causes of Death 43 Chalapati Rao, Alan D. Lopez, and Yusuf Hemed Chapter 6 Population and Mortality after AIDS 59 Rodolfo A. Bulatao Chapter 7 Levels and Patterns of Mortality at INDEPTH Demographic Surveillance Systems 75 Osman A. Sankoh, Pierre Ngom, Samuel J. Clark, Don de Savigny, and Fred Binka Chapter 8 Trends and Issues in Child Undernutrition 87 Todd Benson and Meera Shekar Chapter 9 Diarrheal Diseases 107 Cynthia Boschi-Pinto, Claudio F. Lanata, Walter Mendoza, and Demissie Habte Chapter 10 Developmental Disabilities 125 Geoff Solarsh and Karen J. Hofman Chapter 11 Acute Respiratory Infections 149 Shabir A. Madhi and Keith P. Klugman Chapter 12 Vaccine-Preventable Diseases 163 Mark A. Miller and John T. Sentz v Chapter 13 Tuberculosis 179 Christopher Dye, Anthony D. Harries, Dermot Maher, S. Mehran Hosseini, Wilfred Nkhoma, and Felix M. Salaniponi Chapter 14 Malaria 195 Robert W. Snow and Judy A. Omumbo Chapter 15 Onchocerciasis 215 Uche Amazigo, Mounkaila Noma, Jesse Bump, Bruce Benton, Bernhard Liese, Laurent Yaméogo, Honorat Zouré, and Azodoga Seketeli Chapter 16 Maternal Mortality 223 Khama O. Rogo, John Oucho, and Philip Mwalali Chapter 17 HIV/AIDS 237 Souleymane Mboup, Rosemary Musonda, Fred Mhalu, and Max Essex Chapter 18 Lifestyle and Related Risk Factors for Chronic Diseases 247 Krisela Steyn and Albertino Damasceno Chapter 19 Diabetes Mellitus 267 Jean-Claude Mbanya and Kaushik Ramiaya Chapter 20 Cancers 289 Freddy Sitas, Max Parkin, Zvavahera Chirenje, Lara Stein, Nokuzola Mqoqi, and Henry Wabinga Chapter 21 Cardiovascular Disease 305 Anthony Mbewu and Jean-Claude Mbanya Chapter 22 Mental Health and the Abuse of Alcohol and Controlled Substances 329 Florence K. Baingana, Atalay Alem, and Rachel Jenkins Chapter 23 Neurological Disorders 351 Donald Silberberg and Elly Katabira Chapter 24 Violence and Injuries 361 Brett Bowman, Mohamed Seedat, Norman Duncan, and Olive Kobusingye Index 375 Boxes 1.1 Synopsis of the Key Concerns for DMSSA-1 2 1.2 The Health-Related Millennium Development Goals and Indicators 7 13.1 The WHO Stop TB Strategy 186 13.2 Main Constraints to Improving TB Control, as Identified by National Program Managers in Africa 190 16.1 Model of Three Levels of Delay 233 19.1 Organization of Diabetes Care 282 Figures 1.1 Real GDP per Capita Growth, by Region, 1991­2015 4 1.2 Real GDP per Capita, by Developing Region, 1980­2003 4 1.3 Population below Age 15, 2003 6 1.4 Median Age of Population, 1990 and 2003 6 1.5 Total Fertility Rate, 1990 and 2003 6 1.6 Age Pyramid for Botswana, 2005, with and without AIDS 6 2.1 Infant Mortality Rate in Selected Countries, 1960­2005 13 vi | Contents 2.2 Life Expectancy in Selected Countries 14 3.1 Contrasting Data Availability: The Republic of Congo and Kenya 16 3.2 Estimated Levels of Under-Five and Infant Mortality, by Region of Africa, 2000 18 3.3 Trends in Under-Five Mortality, by Five-Year Period and Region, 1960­2000 19 3.4 Annual Rate of Change in Under-Five Mortality, by Five-Year Period, 1965­2000 20 3.5 Relationships between CHMR and IMR, by Country 20 3.6 Estimated Relationships between CHMR and IMR, by Region 21 3A.1 Fitted Trends Using Different Numbers of Knot Parameters 23 3B.1 Examples of "Quality" Categories for Selected Countries 29 4.1 Probabilities of Dying between Exact Ages 15 and 60 in Sub-Saharan Africa, by Sex 35 4.2 Conditional Probabilities of Dying between Ages 30 and 65 in Zimbabwe, by Sex, from Different Data Sources, 1969­90 36 4.3 Conditional Probabilities of Dying between Ages 15 and 60 in South Africa, by Sex, from Different Data Sources, 1980­2000 36 4.4 Age Distribution of Reported Adult Deaths on the National Population Register of South Africa, by Sex, 1998­2003 37 4.5 Trend in the Estimates of the Probabilities of Dying between Ages 15 and 60 (45q15), by Sex and WHO Region, from DHS Sibling Histories 38 5.1 Global Mortality Strata for GBD 2000 Regions 48 5.2 Estimated Age-Specific Death Rates for Zambia, 2000 49 5.3 Model-Based Predictions of GBD Cause­Group Composition of Mortality by Age and Sex, Zambia, 2000 49 5.4 Summary of GBD Process for Estimating Cause-Specific Mortality in African Countries 53 6.1 Projected Population, Sub-Saharan Africa and Three Selected Countries 62 6.2 Projected Crude Birth Rate, Sub-Saharan Africa and Botswana 64 6.3 Distribution of Women of Reproductive Age in the UN Projection and the No-AIDS Scenario, Botswana 64 6.4 Mortality Effects on Population, Relative to the No-AIDS Scenario, in Different Projections 64 6.5 Projected Life Expectancy, Sub-Saharan Africa and Three Selected Countries 65 6.6 HIV Prevalence and Maximum Loss in Life Expectancy in Alternative Projections 66 6.7 Female Advantage in Life Expectancy, Sub-Saharan Africa and Three Selected Countries 67 6.8 Adult Mortality (45q15) by Sex in UN Projections, Sub-Saharan Africa and Lesotho 68 6.9 Sex Ratios by Age from UN Projections, Zimbabwe 68 6.10 Dependency Ratio, Sub-Saharan Africa and Three Selected Countries 69 6.11 Consistency among Alternative Estimates of Current Adult Mortality (45q15) 70 6.12 HIV Prevalence in 2001 and 2003 70 6.13 Alternative Population Projections, Zambia and Botswana 72 6.14 Alternative Projections of Life Expectancy in Zambia and Botswana 72 7.1 Standard Population Age Structure from INDEPTH, Segi, and WHO 78 7.2 Crude Death Rate and Expectation of Life at Birth 79 7.3 Child Mortality: Probability of Dying between Birth and Age Five (5q0) 80 7.4 INDEPTH Mortality Patterns 1­7, ln(nqx) 83 8.1 The UNICEF Conceptual Framework of the Determinants of Nutritional Status 89 8.2 The Burden of Undernutrition through the Life Cycle and across Generations 91 8.3 Prevalence of Stunting in Children Age 6 to 60 Months, by Country 93 8.4 Subnational Estimates of the Prevalence of Underweight Preschool Children and Area Density of Underweight Children 94 8.5 Stunting Prevalence among Preschoolers, by Urban or Rural Residence, Selected Countries 95 Contents | vii 8.6 Prevalence of Underweight Preschool Children by Wealth Quintile, Selected Countries 95 8.7 Under-Five Mortality Rates and Progress Being Made in Reducing Under-Five Mortality, by Country 96 8.8 Scatter Plot of National Under-Five Mortality and Underweight Prevalence Rates 96 8.9 Daily Dietary Energy Supply Available and Food Production Index, by Country 97 8.10 Access to Safe Water and Adequate Sanitation, by Country 98 8.11 Health Services, by Country 99 8.12 Female Adult Literacy Rates and Girls' Net Primary Enrollment Rates, by Country 100 8.13 Proportion of Infants Receiving Only Breast Milk in the First Year of Life, Selected Countries 101 8.14 Effect of Economic Growth on Attaining the Millennium Development Goal of Reducing the Prevalence of Underweight Preschool Children, Tanzania 105 9.1 Slowing Progress in Child Mortality 108 9.2 Sites with Available Under-Five Diarrhea Mortality Data 109 9.3 Medians of Diarrheal Proportional Mortality among Children under Five in the African Region, 1980­95 111 9.4 Medians of Diarrheal Proportional Mortality among Children under Five in the African Region and Other Developing Regions, 1980­95 111 9.5 Change in Sanitation Coverage by Region, 1990­2000 115 10.1 Relationship between Impairment, Disability, and Handicap (ICIDH) 126 10.2 Relationship between Body Functions, Activities, and Participation (ICF) 126 10.3 Research Steps in the Development of Public Health Interventions 133 10.4 Causal Pathways for Developmental Disabilities 141 12.1 Percentage of Target Population in Africa Vaccinated, by Vaccine Type, 1980­2002 164 12.2 Immunization Coverage with Measles-Containing Vaccines, 2003 166 12.3 Maternal and Neonatal Tetanus Elimination Status, 2002 169 12.4 Global and Regional Immunization Coverage, Three Doses DTP, 1980­2001 169 12.5 Epidemics of Yellow Fever in Africa Reported to the WHO, 1980­2003 170 12.6 African Nations Using Hepatitis B Vaccine in a National Infant Immunization Campaign, 2002 171 12.7 Number of Childhood Vaccines Routinely Used in Developing and Established- Market Countries 173 13.1 Relative Incidence of TB in HIV-Infected Individuals as a Function of CD4 Cell Count 183 13.2 Estimated Incidence Rates of New TB Cases and Percentages of TB Patients Infected with HIV, by Country, 2004 183 13.3 Trends in TB (All Forms) Case Notifications in the WHO African Region Contrasted with Trends in Other Parts of the World 184 13.4 Estimated TB Incidence in Relation to Estimated HIV Prevalence in Adults Age 15 to 49 for 42 Countries in the WHO African Region 184 13.5 Prevalence of HIV in TB Cases (All Forms) in Relation to HIV Prevalence in Adults Age 15 to 49 185 13.6 Progress toward the Target of 70 Percent Case Detection in the WHO African Region, Compared with the Average Progress Worldwide 187 13.7 Expected Reductions in the Number of TB Cases in Kenya over 10 Years 189 13.8 Proportional Reduction in the Incidence of TB over 20 Years among HIV-Positives as a Function of Effective Coverage and the CD4 Count per Microliter at Which People Start ARV Therapy 189 14.1 Malarial Risks in Children Age 0 to 14 Years in a Stable Endemic Area of the Kenyan Coast 196 14.2 Public Health Effects of Plasmodium falciparum Malaria 196 viii | Contents 14.3 Fuzzy Climate Suitability Membership for Malaria 198 14.4 Malaria-Specific and All-Cause Mortality Estimates per Year for Children under Five 207 14.5 Malaria-Specific, Nonmalaria, and All-Cause Mortality Rates among Children under Five 208 15.1 Prevalence of Onchocerciasis Infection in 1974 and Gradual Expansion, 1977­92 218 15.2 Prevalence of Onchocerciasis Infection in 2002 219 15.3 Scaling Up with Additional Interventions 219 16.1 Major Causes of Maternal Mortality in Sub-Saharan Africa 229 16.2 Relation between Skilled Attendant at Delivery and MMR for All Countries, 1995 231 16.3 Effects of the Introduction in Romania of an Anti-Abortion Law in 1966 and Legalization of Abortion in 1989 232 17.1 Disease Burden and Treatment of AIDS in Relation to Global Population and Economy 238 17.2 Global Burden of HIV-1 Infection 238 17.3 Distribution of Major HIV-1 Subtypes and Circulating Recombinant Forms in Africa 241 18.1 Fat and Carbohydrate Intake as Functions of Proportion of Life Spent in a City 251 18.2 Macronutrient Intakes in South Africans, per Capita, 1993 and 1999 251 19.1 Prevalence of Diabetes with Increasing Age in Cameroon 276 19.2 Mean Fasting Blood Glucose by Tertiles of Walking Energy Expenditure in Women: The Cameroon Study 277 19.3 Cost of Different Types of Insulin in Relation to the Gross National Product 282 20.1 Distribution of Radiation Therapy Services in Africa 290 20.2 Major Cancer Types in Sub-Saharan Africa, Both Sexes, All Ages 292 20.3 Number of Cigarettes Consumed per Adult 300 21.1 Cardiovascular Disease, Age-Standardized Rates in the World, 1994­2000 308 21.2 Probability of Death by Broad Cause in Men between the Ages of 15 and 60 Years in Tanzania 317 Tables 1.1 Gross National Income, per Capita, 1980, 1990, 2003 5 1.2 Overview of Health Worker Vacancy Rates for Four Countries 8 1.3 Conflict-Related Deaths by Region 9 2.1 Life Expectancy at Birth for World and UN Regions, 1960­2005 12 2.2 Infant Mortality Rates for World and UN Regions, 1960­2005 13 2.3 Infant Mortality Rates for Sub-Saharan Africa and UN Subregions, 1960­2005 13 2.4 Life Expectancy at Birth for Sub-Saharan Africa and UN Subregions, 1960­2005 13 3B.1 Estimates of Under-Five Mortality, by Country and Year 26 3B.2 Estimates of Infant Mortality, by Country and Year 27 3B.3 Countries and Data Sources 28 4.1 Probabilities of Dying between Ages 15 and 60 (45q15) in Sub-Saharan Africa 34 4.2 Probabilities of Dying between Ages 15 and 60 (45q15), by Sex, according to Intercensal Survival, Malawi 35 4.3 Orphanhood Estimates of Person-Years Lived and Lost at Ages 35­75 (Men) and 25­75 (Women), Kenya 35 4.4 Estimates of the Probability of Dying between Ages 15 and 60(45q15), from DHS Sibling Histories 40 5.1 Proportion of Deaths Due to Diseases and Injuries in Children under Five Years, by Source of Data 44 5.2 Proportion of Deaths Due to Diseases and Injuries in Children Age 5 to 14 Years, by Source of Data 44 Contents | ix 5.3 National Vital Records Data: Proportionate Distribution of Cause of Death of Children under Five 46 5.4 National Vital Records Data: Proportionate Distribution of Cause of Death at Age 5 to 14 Years 46 5.5 National Vital Records Data: Proportionate Distribution of Cause of Death at Age 15 to 44 Years 47 5.6 National Vital Records Data: Proportionate Distribution of Cause of Death at 45 Years of Age and Older 47 5.7 Epidemiological Estimates of Malaria Mortality in Nonpregnant African Population, 1995 50 5.8 Estimates of Diarrheal Deaths of Children under Five, 2000 50 5.9 GBD Estimates of Leading Causes of Death, by Sex, 2000 53 5.10 GBD Estimates of Leading Causes of Death in AFR D and AFR E, 2000 54 5.11 GBD Estimates of Leading Causes of Death at Age 0 to 14 Years, 2000 55 5.12 GBD Estimates of Leading Causes of Death at Age 15 to 59 Years, Subregional Comparison, 2000 55 5.13 GBD Estimates of Leading Causes of Death in AFR E at Age 15 to 59 Years: Comparison between Males and Females, 2000 56 5.14 GBD Estimates of Leading Causes of Death at Age 60 Years and Older: Comparison between Males and Females, 2000 56 6.1 Effects of Demographic Factors in Producing Differences between Projections 63 6.2 Estimates of HIV Prevalence among Adults Age 15 to 49, 2001­03 71 7.1 Summary of Mortality Data from INDEPTH Sites, 1995­99 77 7.2 Crude Death Rates and Expectation of Life at Birth 79 7.3 Infant and Child Mortality 80 7.4 Adult Mortality and Child and Adult Mortality Ratio 81 8.1 Percentage of Total Estimated Annual Burden of Disease in Africa Attributed to Major Risk Factors 88 8.2 Indicators of Nutritional Status 90 8.3 Effects of Deficiency of Micronutrients and Principal Dietary Sources 92 8.4 Projected Impact of Economic Growth and Direct Large-Scale Nutrition Intervention Programs on Stunting and Poverty, Kagera Region, Tanzania, 2015 102 9.1 Median Estimates of Episodes of Diarrhea per Child per Year in the African Region, by Age Group 110 9.2 Estimated Number and Proportion of Deaths Due to Diarrhea among Children under Five 112 9.3 Available Etiology Data, by Country and Study Setting 113 9.4 Median of the Proportions of Etiological Agents among Children under Five in the AFR D Subregion, by Study Site 113 9.5 Median of the Proportions of Etiological Agents among Children under Five in the AFR E Subregion, by Study Site 114 9.6 Feces Disposal Practices at Home, by Urban and Rural Residences in Four African Countries 115 9.7 Prevalence (Median) of Diarrheal Disease, the Two Weeks before Survey, by Urban and Rural Site of Residence in AFR D and AFR E 116 9.8 Children under Five Taken to Health Facilities or Receiving Treatment for Diarrheal Disease in the Two Weeks before Survey, by Sex 116 9.9 Cases and Deaths Due to V. cholerae Reported to WHO, 1996­2001 116 9A.1 Regional Reporting Categories for Global Burden of Disease 2000: WHO African Subregions 117 9A.2 Main Characteristics of the Studies Included in the Morbidity Review 117 9A.3 Main Characteristics of the Studies Included in the Mortality Review 118 x | Contents 9A.4 Main Characteristics of the Studies Included in the Etiology Review 119 10.1 Prevalence of Disability in Sub-Saharan Africa 129 10.2 Prevalence of Disability in Other Developing Countries 130 10.3 Prevalence of Hearing Disability in Sub-Saharan Africa 131 10.4 Estimated Birth Prevalence of Infants with Serious Congenital Disorders, by WHO Region 133 11.1 Results of Lung Aspirations of Children Who Had Not Received Antibiotics and Not Mentioned Underlying Illness 154 11.2 Estimated Incidence of Organism-Specific, Bacteremic, Community-Acquired Severe LRTI in HIV-1-Infected and HIV-1-Uninfected Children Age 2 to 24 Months 154 11.3 Estimated Incidence for Specific Viral-Associated Severe LRTI in HIV-1-Infected and HIV-1-Uninfected Children Age 2 to 23 Months 155 12.1 WHO-Estimated Deaths and DALYs from Vaccine-Preventable Diseases, 2002 164 12.2 Potential Deaths Averted by HBV, Hib, Rotavirus, and SP Vaccine Implementation 174 13.1 The Contribution of Sub-Saharan Africa to the Global TB Epidemic 184 14.1 Populations at Risk During 2000 199 14.2 Median and Interquartile Ranges of Malaria-Specific Mortality Estimates per 1,000 People per Year 201 14.3 Estimated Numbers of Malaria-Specific Deaths and Interquartile Ranges during 2000 202 15.1 Scaling up APOC, 1996­2003 220 16.1 Maternal Mortality Measures, 1990, 1995, and 2000 226 16.2 Maternal Mortality Measures in Sub-Saharan Africa, by Country, 1990­2000 227 17.1 HIV-1 Prevalence in Representative Regions and Countries 239 18.1 Exposure to Tobacco Products of Participants, Age 13­15 Years, in the Global Youth Tobacco Survey, 1999­2001 249 18.2 Prevalence of Adults Who Smoke Cigarettes, by Country 250 18.3 Anthropometric Indicators, by Country 254 18.4 Large Prevalence Studies on Hypertension since 1997, by Country 256 18.5 Mean Lipid Levels and Prevalence of Dyslipidemia in Black Subjects 259 18.6 The Clustering of NCD Risk Factors in Nigerian Subjects 260 19.1 Data Sources for the Prevalence of Type 2 Diabetes and IGT, by Year of Study 268 19.2 Data Sources for Prevalence Estimates of Diabetes Mellitus and IGT, by Country, 2003 269 19.3 Data Sources for the Prevalence of Diabetes Complications, by Disease and Year 271 19.4 Prevalence Estimates of Diabetes Mellitus, by Country, 2003 274 19.5 Prevalence Estimates of IGT, by Country, 2003 275 19.6 Projections of Diabetes and IGT from 2003 to 2025 in the Age Group of 20 to 79 Years 276 19.7 Calculated Estimates of the Costs of Diabetes Care, by Country 281 20.1 Estimated Percentages of Deaths, by Cause, 2002 290 20.2 Cumulative Incidence of Cancer in Women up to 64 Years of Age, 1993­97 290 20.3 Estimated Number of New Cases and Age-Standardized (World) Incidence Rates for the Leading Cancers in Males and Females, 2002 291 20.4 Prevalence of Hepatitis C Virus IgG Antibodies in Sub-Saharan Africa, 2000 296 21.1 The Epidemiological Transition in the Seychelles, 1976 and 1994 306 21.2 Incidence of Stroke 310 21.3 The Relation between CVD, Risk Factors, Behaviors, and Determinants 314 21.4 Risk Factors for CHD Reported from Hospital Patients 315 21.5 Prevalence of Hypertension, by Country 316 21.6 Prevalence of Diabetes, by Country 318 21.7 Prevalence of Tobacco Smoking, by Country 318 Contents | xi 21.8 Prevalence of Alcohol Consumption, by Country 318 21.9 Prevalence of Obesity, by Country 319 21.10 Prevalence of Obesity in Women Age 15 to 49 Years, by Country 319 22.1 Age-Standardized Incidence, Prevalence, and Mortality Rate Estimates for WHO AFRO Epidemiological Subregions, 2000 332 22.2 YLD, YLL, and DALY Estimates for WHO AFRO Epidemiological Subregions, 2000 333 22.3 Summaries of Selected Studies on Suicide and Parasuicide 336 22.4 Selected Studies on the Psychosocial and Mental Health Consequences of HIV/AIDS 340 22A.1 Selected Sub-Saharan Africa Data on Mental Health Disorders 345 24.1 Cause of Injury by Outcome: Mukono District, Uganda 366 24.2 Cause of Injury by Outcome: Kawempe Division, Kampala District, Uganda 366 24.3 Cause of Injury by Outcome: Gulu District, Uganda 367 24.4 Top Three Causes of Injury, by Age, Rural District (Mukono), Uganda 367 24.5 Top Three Causes of Injury, by Age, Urban Division (Kawempe), Uganda 368 xii | Contents Foreword More attention is now focused on improving the health of the successes, a common theme in these and other successful pro- population of Sub-Saharan Africa than at any previous time. In grams has been the emphasis on the monitoring of disease indi- the years since the publication of the first edition of Disease and cators and the effectiveness of programs to address them. Mortality in Sub-Saharan Africa in 1991, numerous reports have Without knowledge of disease incidence, prevalence, and sever- been issued by national governments, development agencies, ity, setting policies for prioritizing interventions risks misallo- and researchers addressing the health status of African popula- cating resources to combat causes of ill health that contribute tions and proposing strategies to more effectively combat poor little to the overall health status of a population. Good epidemi- health with improved delivery of health services to prevent and ological information does not ensure good policy decisions or cure diseases. Prime among these was the World Bank's 2005 effective implementation, but without reliable epidemiological report Improving Health, Nutrition and Population Outcomes in data, efforts to design cost-effective strategies and to implement Sub-Saharan Africa--The Role of the World Bank, which gave technologies are no more than theoretical exercises. rise to the publication of this book. Increased funding for health Since the publication of the first edition of Disease and from governments, multilateral and bilateral donors, as well as Mortality in Sub-Saharan Africa, many new sources of health new public-private partnerships and foundations has become and demographic information have become available, includ- available for assisting African countries to deliver more effective ing data on trends in HIV infection from antenatal clinic sur- health interventions. The Millennium Development Goals have veillance sites, the first set of African life tables from a growing focused the attention of the world on achieving a clear set of number of demographic surveillance sites, injury statistics goals--several of which are directly concerned with improving from a small number of injury mortality surveillance registers, health outcomes--to be achieved by 2015. and cancer data from cancer registers. Improved methods for Yet the sobering reality is that life expectancy has decreased estimating the incidence of several other diseases, including by almost five years for the continent as a whole since the 1991 tuberculosis, maternal mortality, and chronic diseases, have publication, and by much more in some countries. As the chap- also improved the reliability of health statistics. Verbal autopsy ters in this volume document, children under five are dying at studies have linked with demographic surveillance sites, adding unacceptably high rates from causes for which effective inter- to our knowledge on changes in the cause-of-death composi- ventions exist, and adult mortality from infectious diseases tion in several countries. has risen to extraordinary levels. HIV/AIDS has spread from Notwithstanding these advances in health statistics, a theme eastern Africa to the rest of the continent, affecting southern that emerges from all the chapters in this volume is that too African countries the most. Malaria mortality of children little is known about trends in the diseases and conditions increased during the 1990s, and TB has reemerged as a leading included here in order to monitor and evaluate the effective- cause of death for adults, largely due to the spread of AIDS. Not ness of programs intended to produce better health outcomes. surprisingly, at this time Sub-Saharan Africa is not on track to As we get closer to the 2015 end point of the Millennium reach any of the health Millennium Development Goals. Development Goals, reaching the goals will become increas- It is important to recognize that not all trends have been neg- ingly challenging. The continued improvement of disease sur- ative. The prevalence of HIV/AIDS has significantly decreased veillance and other regularly published health information in several African countries, including Uganda, one of the remains as important a priority for African health systems as it worst-affected countries at the time of the publication of the was for the first edition. first edition. Measles mortality has been virtually eliminated in Callisto Madavo the countries of southern Africa in the past decade. Enormous Former Vice-President strides continued to be made in the control of onchocerciasis Africa Region during the 1990s. Although many factors contributed to these The World Bank xiii Acknowledgments This publication consists of the contributions of 70 authors, School of Medicine, UK), Robert Redfield (University of coordinated by a group of editors at the World Bank (Florence Maryland, USA), Brian Robertson (Cape Town University, Baingana, Eduard Bos, and Khama Rogo), the U.S. National South Africa), Daniel M. Sala-Diakanda (IFORD, Cameroon), Institutes of Health (Karen Hofman and Dean Jamison), the Eugene Sobngwi (Yaounde University, Cameroon), David Global Fund to Fight AIDS, Tuberculosis, and Malaria (Richard Thomas (Fred Hutchinson Cancer Research Center, USA), and Feachem), and the South African Medical Research Council Mark Wainberg (McGill University, Canada). (Malegapuru Makgoba). Management of the publication was The publication of this volume was made possible by the carried out jointly at the World Bank and, for a subset of chap- generous support of the government of the Netherlands ters, at the South African Medical Research Council. through the Bank-Netherlands Partnership Program (BNPP) The editors are grateful to the chapter authors, named at the at the World Bank. The editors are grateful to Julie McLaughlin beginning of each of their respective chapters, who worked and Ok Pannenborg of the Africa Region of the World Bank tirelessly to produce the various drafts and revisions. for initiating and subsequently overseeing the whole process The editors thank the following reviewers of draft chapters: leading to the publication of this book. Administrative and Larry Barat (World Bank, USA), George Bicego (South Africa logistical support was provided by Carole Roberts of the South CDC, South Africa), Gretchen Birbeck (Michigan State African Medical Research Council and Richard Babumba of University, USA, and The Gambia), Martien Borgdoff (The the World Bank. Rifat Hasan of the World Bank guided the Netherlands), Julie Cliff (Mozambique), Jerry Coovadiah manuscript through the final stages and produced an (University of Natal, South Africa), Andrew Grulich "Executive Summary" for distribution at the High-Level (University of New South Wales, Australia), David Gwatkin Forum on the Health Millennium Development Goals meeting (World Bank, USA), Kenneth Hill (Johns Hopkins University, in Paris, 2005. Anne-Sophie Ville and Willyanne DeCormier USA), Adnan Hyder (Johns Hopkins University, USA), Jean- Plosky also provided support. Claude Mbanya (Yaounde University, Cameroon), Samuel The editors are also grateful to the World Bank and the Lantei Mills (World Bank, USA), Pindile Mntla (Medical Fogarty International Center of the U.S. National Institutes of University of Southern Africa, South Africa), David Ndetei Health, which allowed the editors and authors from these insti- (Nairobi University, Kenya), Steven Obaro (Imperial College tutions to dedicate staff time to contribute to this publication. xv Contributors Volume Editors Bruce Benton, Public Health Adviser, Human Development Department, Africa Region, The World Bank, Dean T. Jamison, Professor, Institute for Global Health, Washington, DC University of California, San Francisco Fred Binka, Executive Director, INDEPTH Network, Accra, Richard G. Feachem, Executive Director, Global Fund to Ghana Fight AIDS, Tuberculosis, and Malaria; and Director, Institute for Global Health, University of California, San Francisco Eduard R. Bos, Lead Population Specialist, Human and Berkeley Development Network, The World Bank, Washington, DC Malegapuru W. Makgoba, Vice-Chancellor and Principal, Cynthia Boschi-Pinto, Medical Officer, Department of Child University of KwaZulu-Natal, South Africa and Adolescent Health and Development, World Health Organization, Geneva Eduard R. Bos, Lead Population Specialist, Human Development Network, The World Bank, Washington, DC Brett Bowman, Senior Researcher, Institute for Social and Health Sciences, University of South Africa Florence K. Baingana, Senior Health Specialist, Human Development Network, The World Bank, Washington, DC Debbie Bradshaw, Director, Burden of Disease Research Unit, Medical Research Council, South Africa Karen J. Hofman, Director, Division of Advanced Studies and Policy Analysis, Fogarty International Center, National Rodolfo A. Bulatao, Independent Consultant, Washington, DC Institutes of Health, Washington, DC Jesse Bump, Consultant, Onchocerciasis Coordination Unit, Khama O. Rogo, Lead Specialist, Africa Region, The World Human Development Department, Africa Region, The World Bank, Washington, DC Bank, Washington, DC Zvavahera Chirenje, Lecturer/Consultant, Department of Chapter Authors Obstetrics and Gynaecology, College of Health Sciences, Jacob Adetunji, Technical Adviser, U.S. Agency for University of Zimbabwe, Harare International Development, Washington, DC Samuel J. Clark, Assistant Professor, Department of Atalay Alem, Psychiatrist, Amanuel Psychiatric Hospital, Sociology, University of Washington, Seattle; Research Faculty of Medicine, Addis Ababa University, Ethiopia Associate, Institute of Behavioral Science, University of Colorado at Boulder; and Research Officer, MRC/Wits Rural Uche Amazigo, Director, African Programme for Public Health and Health Transitions Research Unit Onchocerciasis Control, Ouagadougou, Burkina Faso (Agincourt), School of Public Health, University of the Witwatersrand, South Africa Agbessi Amouzou, PhD Candidate, Department of Population Dynamics, John Hopkins University, Baltimore, Maryland Albertino Damasceno, Professor of Cardiology, Faculty of Medicine, Eduardo Mondlane University, Mozambique Florence K. Baingana, Senior Health Specialist, The World Bank, Washington, DC Don de Savigny, Head of Unit, Clinical and Intervention Epidemiology, Swiss Tropical Institute, Basel, Switzerland Todd Benson, Research Fellow, Food Consumption and Nutrition Division, International Food Policy Research Norman Duncan, Chair, Department of Psychology, Institute, Washington, DC University of the Witwatersrand, South Africa xvii Christopher Dye, Coordinator, Tuberculosis Monitoring and Shabir A. Madhi, Codirector, Medical Research Council Evaluation, Stop TB Department, World Health Organization, Respiratory and Meningeal Pathogens Research Unit, National Geneva Institute of Communicable Diseases, University of the Witwatersrand, Johannesburg, South Africa Max Essex, Chair, Department of Immunology and Infectious Diseases, Harvard AIDS Institute, and the Botswana-Harvard Dermot Maher, Medical Officer, Stop TB Department, World AIDS Institute Partnership, Harvard School of Public Health, Health Organization, Geneva Boston, Massachusetts Jean-Claude Mbanya, Endocrine and Diabetes Unit, Demissie Habte, International Director, James P. Grant Department of Internal Medicine and Specialities, Faculty of School of Public Health, BRAC University, Bangladesh Medicine and Biomedical Sciences, University of Yaoundé I, Yaoundé, Cameroon Anthony D. Harries, Technical Adviser, HIV Care and Support, Ministry of Health, Lilongwe, Malawi Anthony Mbewu, President, Medical Research Council, South Africa; and Visiting Professor in Medicine and Cardiology, Yusuf Hemed, Coordinator, MEASURE Evaluation, Dar es University of Cape Town, South Africa Salaam, Tanzania Souleymane Mboup, Professor of Microbiology, Laboratory Kenneth Hill, Professor, Department of Population and of Bacteriology and Virology, CHU Le Dantec, Université Family Health Sciences, Johns Hopkins University, Baltimore, Cheikh Anta Diop, Dakar, Senegal Maryland Walter Mendoza, Researcher, Instituto de Investigación Karen J. Hofman, Director, Division of Advanced Studies and Nutricional, Lima, Peru Policy Analysis, Fogarty International Center, National Institutes of Health, Bethesda, Maryland Fred Mhalu, Professor of Microbiology and Immunology, Muhimbili University College of Health Sciences, Dar es S. Mehran Hosseini, Epidemiologist, Stop TB Department, Salaam, Tanzania World Health Organization, Geneva Mark A. Miller, Director, Division of International Rachel Jenkins, Director, World Health Organization­United Epidemiology and Population Studies, Fogarty International Kingdom Collaborating Centre, Institute of Psychiatry, Kings Center, National Institutes of Health, Bethesda, Maryland College London Nokuzola Mqoqi, National Cancer Registry and Cancer Elly Katabira, Neurologist, Department of Neurology, Epidemiology Research Group, National Health Laboratory Makerere Medical School, Kampala, Uganda. Service, Johannesburg, South Africa Keith P. Klugman, Codirector, Medical Research Council Rosemary Musonda, Director, National AIDS Council, Respiratory and Meningeal Pathogens Research Unit, National Lusaka, Zambia Institute of Communicable Diseases, University of the Philip Mwalali, International Health Consultant and Medical Witwatersrand, Johannesburg, South Africa; and Professor, Adviser, African Economic Foundation, Los Angeles, Department of Global Health, Rollins School of Public Health California and Division of Infectious Diseases, School of Medicine, Emory University, Atlanta, Georgia Pierre Ngom, Senior Research Adviser, Family Health International, Nairobi, Kenya Olive Kobusingye, Adviser, Violence and Injury Prevention, and Disabilities, World Health Organization Regional Office Wilfred Nkhoma, Regional Adviser (Tuberculosis), World for Africa, Brazzaville, Congo Health Organization, Harare, Zimbabwe Claudio F. Lanata, Senior Researcher, Instituto de Investigación Mounkaila Noma, Chief, Epidemiology and Vector Nutricional, Lima, Peru Elimination Unit, African Programme for Onchocerciasis Control, Ouagadougou, Burkina Faso Bernhard Liese, Public Health Adviser, Human Development Department, Africa Region, The World Bank, Judy A. Omumbo, Research Fellow, Public Health Group, Washington, DC Kenya Medical Research Institute-Wellcome Trust Collaborative Program, Nairobi, Kenya Alan D. Lopez, Professor and Head of School, School of Population Health, University of Queensland, Herston, John Oucho, Professor of Demography and Chairman, Australia African Population and Environment Institute, Nairobi, Kenya xviii | Contributors Max Parkin, Chief, Unit of Descriptive Epidemiology, Freddy Sitas, Director, Cancer Research and Registers International Agency for Research on Cancer, Lyons, France Division, The Cancer Council of New South Wales, Australia Kaushik Ramiaya, Consultant Physician and Assistant Medical Administrator, Shree Hindu Mandai Hospital, Robert W. Snow, Professor, Tropical Public Health, Centre for Dar es Salaam, Tanzania Tropical Medicine, University of Oxford, United Kingdom; and Head, Public Health Group, Kenya Medical Research Chalapati Rao, Lecturer, School of Population Health, Institute-Wellcome Trust Collaborative Program, Nairobi, University of Queensland, Herston, Australia Kenya Khama O. Rogo, Lead Specialist, Africa Region, The World Geoff Solarsh, Professor and Head of School, Monash Bank, Washington, DC University School of Rural Health, Bendigo, Australia Felix M. Salaniponi, Programme Director, National Lara Stein, Acting Director, National Cancer Registry and Tuberculosis Control Programme, Ministry of Health, Cancer Epidemiology Research Group, National Health Lilongwe, Malawi Laboratory Service, Johannesburg, South Africa Osman A. Sankoh, Manager, Communications and External Krisela Steyn, Director, Chronic Diseases of Lifestyle Unit, Relations, INDEPTH Network Secretariat, Accra, Ghana Medical Research Council, South Africa Mohamed Seedat, Director, Institute for Social and Health Ian M. Timaeus, Head, Centre for Population Studies and Sciences, University of South Africa, and South African Professor of Demography, London School of Hygiene and Medical Research Council-University of South Africa Crime, Tropical Medicine Violence and Injury Lead Programme Henry Wabinga, Professor, Kampala Cancer Registry, Azodoga Seketeli, Medical Entomologist, Former Director, Department of Pathology, Makerere University, Kampala, African Programme for Onchocerciasis Control, Uganda Ouagadougou, Burkina Faso Laurent Yaméogo, Coordinator, Office of Programme John T. Sentz, Research Assistant, Division of International Director, African Programme for Onchocerciasis Control, Epidemiology and Population Studies, Fogarty International Ouagadougou, Burkina Faso Center, National Institutes of Health, Bethesda, Maryland Honorat Zouré, Biostatistic and Mapping, African Meera Shekar, Senior Nutrition Specialist, Human Programme for Onchocerciasis Control, Ouagadougou, Development Network, The World Bank, Washington, DC Burkina Faso Donald Silberberg, Professor of Neurology, University of Pennsylvania, Philadelphia Contributors | xix Abbreviations and Acronyms ADR adverse drug reactions DMSSA-1 Disease and Mortality in Sub-Saharan Africa, AFB acid-fast bacilli first edition AIDS acquired immune deficiency syndrome DMSSA-2 Disease and Mortality in Sub-Saharan Africa, AMMP Adult Morbidity and Mortality Project second edition anti-GAD glutamic acid decarboxylase antibodies DOTS directly observed treatment, short course APOC African Programme for Onchocerciasis DSS demographic surveillance system Control DTP diphtheria, tetanus, pertussis AR androgen receptor EAEC entero-adherent pathogenic Escherichia coli ARI acute respiratory infections EIR entomological inoculation rates, ARV antiretroviral EMF endomyocardial fibrosis AUDIT Alcohol Use Disorders Identification Test EmOC emergency obstetric care BCG bacillus Calmette-Guérin EPI Expanded Program on Immunization BMI body mass index EPTB extrapulmonary tuberculosis BOMA Burden of Malaria in Africa ETEC enterotoxigenic Escherichia coli BOSTID Board of Science and Technology for FAO Food and Agriculture Organization of the International Development United Nations CBR community-based rehabilitation FCS fuzzy climate suitability CDTI community-directed treatment with FWCW Fourth World Conference on Women ivermectin GAVI Global Alliance for Vaccines and Immunization CHD coronary heart disease GHQ general health questionnaire CHMR child mortality rate GIS Geographical Information System CI confidence interval GBD global burden of disease CIESIN Center for International Earth Science GDM gestational diabetes mellitus Information Network GDP gross domestic product CIN cervical intraepithelial neoplasia GHQ General Health Questionnaire CM cerebral malaria HAART highly active antiretroviral therapy CMD common mental disorders HAV hepatitis A virus CMO chronic otitis media HBIG hepatitis B immunoglobin CNS central nervous system HBV hepatitis B CRF circulating recombinant forms HES hypereosinophilic syndrome CRS congenital rubella syndrome HHV8 human herpesvirus-8 CSO Central Statistical Office Hib Haemophilus influenzae type B CTL cytolytic T cell HIC high-income country CVD cardiovascular disease HIPC heavily indebted poor country CWIQ Core Welfare Indicators Questionnaire HIV human immunodeficieny virus DALY disability-adjusted life year HMIS health management information systems DCCT Diabetes Control and Complication Trial HPV human papillomavirus DCM dilated cardiomyopathy HSCL Hopkins Symptom Checklist DHS Demographic and Health Survey ICA islet cell antibodies xxi ICD-10 International Statistical Classification of Diseases PA-ICPD Programme of Action of the International and Related Health Problems, 10th revision Conference on Population and Development IDD iodine deficiency disorder PCP Pneumocystis carinii pneumonia IFG impaired fasting glycemia PEM protein-energy malnutrition IGT impaired glucose tolerance PMDF proportion of deaths of women of reproductive IHD ischemic heart disease ages due to maternal causes IMR infant mortality rate PPP international dollars INCLEN International Clinical Epidemiology Network PRSP poverty reduction strategy paper IPD invasive pneumococcal disease PSE Present State Examination IPT isoniazid preventive treatment PTB pulmonary tuberculosis IPTT Initiative for Technology Transfer PTSD posttraumatic stress disorder IPV inactivated poliovirus vaccine PYO person-years of observation IQI interquartile intervals RAMOS Reproductive Age Mortality Studies IQR interquartile ranges REMO Rapid Epidemiological Mapping of IUATLD International Union against TB and Lung Disease Onchocerciasis IUGR intrauterine growth retardation RHD rheumatic heart disease IVIG intravenous immunoglobulin RR relative risk LBW low birthweight RTI road traffic injury LMIC low- to middle-income country SADC Southern African Development Community LRTI lower respiratory tract infections SAVVY Sample Vital Registration and Verbal Autopsy MDG Millennium Development Goal SIR Susceptible-Infected-Removed MDR multidrug resistance SIV simian immunodeficiency virus MICS Multiple Indicator Cluster Surveys SMA severe malarial anemia MMR maternal mortality ratio SRQ-25 Self Report Questionnaire--25 item version MOTT mycobacteria other than tuberculosis SSRI selective serotonin reuptake inhibitor MR mental retardation STD sexually transmitted disease NCD noncommunicable disease TB tuberculosis NDS National Demographic Survey TCA tricyclic antidepressant NEPAD New Partnership for Africa's Development TFR total fertility rate NGDO nongovernmental development organization TMP-SMX trimethoprim-sulfamethoxazole NGO nongovernmental organization TT tetanus toxoid NIMSS National Injury Mortality Surveillance System U5MR under-five mortality rate NNRTI nonnucleoside reverse transcriptase inhibitor UKPDS United Kingdom Prospective Diabetes Study NNT neonatal tetanus UNAIDS Joint United Nations Programme on HIV/AIDS NRTI nucleoside analogue reverse transcriptase UNDP United Nations Development Programme inhibitor UNICEF United Nations Children's Fund NVP nevirapine URTI upper respiratory tract infection OCP Onchocerciasis Control Programme VA verbal autopsy OEPA Onchocerciasis Elimination Program for the VAD vitamin A deficiency Americas VAPP vaccine-associated paralytic poliomyelitis OPV oral poliovirus vaccine VR vital statistics registration system OR odds ratio WFS World Fertility Survey ORT oral rehydration therapy WHO World Health Organization OSD onchocercal skin disease YLD years lived with disability xxii | Abbreviations and Acronyms Chapter 1 Changing Patterns of Disease and Mortality in Sub-Saharan Africa: An Overview Florence K. Baingana and Eduard R. Bos Fifteen years have passed since the first edition of Disease box 1.1. The most significant impact on disease and mortal- and Mortality in Sub-Saharan Africa (DMSSA-1) was pub- ity in Africa has been the growth of the HIV/AIDS epidemic, lished. Its main purpose was to assist the World Bank's work which has infected more than 30 percent of adults in some in the health sector by describing conditions and diseases countries while spreading across the continent. Its impact that contributed most to the overall burden of disease and has changed trends in many of the diseases covered in this by identifying ways to prevent and manage these causes of ill volume and dramatically worsened the overall level of mor- health. The volume was timely because of the adverse effect tality in many African countries. The potential impact of the economic downturn of the early 1980s had on health in HIV/AIDS was anticipated in DMSSA-1; the current volume Africa and because of the need to evaluate the impact of pri- documents the burden the disease is currently inflicting on mary health care strategies that had been promoted in the Africa. preceding decade. Epidemiologic information coming from demographic surveillance sites that had not previously been fully compared and disseminated provided a new source for APPROACH assessing trends in mortality. All this occurred against a backdrop of increasing concern about how the human Although the second edition (hereafter called DMSSA-2) immunodeficiency virus/acquired immune deficiency syn- has the same overall objective of informing policy makers drome (HIV/AIDS), then still a relatively new and geo- (at the World Bank as well as in countries and among other graphically more limited disease, could potentially affect development partners), the approach taken to compile the health and development in Africa. information was quite different from that for the first edi- In the years since the publication of DMSSA-1 in 1991, tion. DMSSA-1 was organized in three broad sections, cov- epidemiological and demographic changes have occurred ering patterns of mortality, diseases and conditions, and that require an update if the volume is to remain useful for longitudinal studies of mortality in demographic surveil- policy makers in addressing the "Key Concerns" shown in lance sites. In DMSSA-2, the number of chapters covering 1 Box 1.1 Synopsis of the Key Concerns for DMSSA-1 "What are, or should be, the information needs of policy problems of differing types are most severe, which con- makers? How can available analyses and data be best ditions and diseases are placing the greatest burden on presented to serve those needs? How can the methods of the populations (and on the health care system), why data collection and analysis that are now available diseases occur (through quantification of risk factors), improve the information base for policy?" and how most efficiently and effectively to prevent dis- Policy makers must address the questions of whether ease and manage those cases that do occur. the health sector ought to be a priority concern, where Source: Feachem and Jamison 1991. diseases and conditions has been expanded from 8 to 17 Lifestyle and Related Risk Factors for NCDs. Increased use (out of a total of 24 chapters), with greater emphasis on of tobacco and increased consumption of fats, sugar, alco- emerging noncommunicable conditions and injuries. The hol, and animal products are critical risk factors for many section discussing the demographic surveillance sites has NCDs. At the same time, the amount of physical exercise has been dropped, and the information from the sites is now been decreasing, leading to a sedentary lifestyle that is asso- covered in a synthesis chapter that enables a better compar- ciated with obesity, diabetes, and hypertension. This chapter ative perspective. The number of authors and editors has provides an overview of the risk factors for the NCDs dis- increased along with the number of chapters: there are now cussed in subsequent chapters. 24 chapters with one to eight authors each (for a total of 70); Diabetes Mellitus. Three million people in Sub-Saharan most chapters have at least one author from Sub-Saharan Africa were afflicted with type 2 diabetes as of 1994, but that Africa. number is projected to increase by two- or threefold by 2010. The highest prevalence is found among populations of Indian descent, urban populations, and those with a family history of diabetes, obesity, or physical inactivity. The chap- CONDITIONS NOT COVERED IN DMSSA-1 ter includes a discussion of studies of diabetes onset and mortality in Tanzania and Zimbabwe. Challenges to the pro- DMSSA-1 emphasized communicable diseases, which are vision of health care for diabetes in Sub-Saharan Africa responsible for the largest disease burden and cause the include short consultation times, inadequately trained staff, highest number of deaths. The burden of communicable nonexistent referral systems, inadequate levels of staff, and diseases has increased since the publication of the first poor record keeping. edition, largely owing to the rapid rise in HIV/AIDS. Non- Cancers. Cancers have been a low priority in Sub-Saharan communicable diseases, however, are also becoming a sig- Africa, yet the probability of a 65-year-old woman develop- nificant burden in several countries, leading to dual burdens ing cancer in Sub-Saharan Africa is only 20 percent lower of disease. DMSSA-1 combined cardiovascular disease and than in Western Europe. Factors affecting cancer incidence cancers in one chapter; DMSSA-2 expands the coverage of and mortality include increases in the prevalence of tobacco noncommunicable diseases (NCDs) substantially. Chapters consumption; HIV-induced immunosuppression; increased on the following diseases and conditions have been added: use of alcohol; the high prevalence of cancer-associated Developmental Disorders. This chapter discusses the high- agents like papilloma viruses, hepatitis B virus, and human er rates of severe mental retardation, visual impairment, and herpes virus 8; and exposure to aflatoxins. The top three hearing impairment found in Sub-Saharan Africa than in cancers for men are Kaposi's sarcoma, liver cancer, and more developed regions. An estimated 47 percent of visual prostate cancers; for women, cervical cancer, breast cancer, and 50 to 66 percent of hearing impairments in Sub- and Kaposi's sarcoma. Saharan Africa are found to be preventable. Risk factors Cardiovascular Diseases. Cardiovascular disorders are the include congenital disorders, perinatal and neonatal condi- second most common cause of adult deaths in Sub-Saharan tions, infections, environmental toxins, accidents, injuries, Africa, as well as a major cause of chronic illness and dis- and malnutrition. ability. Half of cardiovascular disease (CVD) deaths occur 2 | Florence K. Baingana and Eduard R. Bos among people 30 to 69 years of age, which is 10 or more and demographic developments, as well as the changes in years younger than in more developed regions. Incidence of how health in Africa is addressed by development stroke in Sub-Saharan Africa is estimated to be about 1 per organizations. 1,000. Survival outcomes are poor, due to delayed hospital- ization, absence of thrombolysis and angioplasty, and low socioeconomic status and illiteracy. Rheumatic heart dis- The Impact of HIV/AIDS ease, still prevalent among children and teenagers, is a disease of poverty that is related to overcrowding, poor A striking feature of DMSSA-2 is the documentation of the housing, and undernutrition. direct impact of HIV/AIDS on the epidemiology of almost Mental Health, Alcohol and Substance Abuse. Depression all infectious diseases included in this volume, as well as on in Sub-Saharan Africa is estimated to have an incidence rate overall adult and child mortality. According to the United of 15 to 18 percent and a lifetime prevalence rate of 18 to Nations' (UN) 2004 projections, life expectancy at birth has 30 percent. Common mental disorders (depression and anx- dropped by three years since 1990 for the region as a whole; iety) have a point prevalence rate that ranges from 1 to for countries most affected by HIV/AIDS, the drop in life 5 percent. The point prevalence rate for schizophrenia is expectancy has been 20 years or more. the same as in other parts of the world, ranging from 2 to As shown in the chapters in this volume, HIV is linked to 5 per 1,000 population, with a lifetime prevalence of 7 to worsening trends in many diseases, for both adults and chil- 9 per 1,000. The Sub-Saharan Africa region, the most dren. For example, Madhi and Klugman (chapter 11) state conflict-affected region of the world, has seen rates of post- that as much as 45 percent of hospitalizations and 80 per- traumatic stress disorder (PTSD), anxiety, and depression cent of deaths due to lower respiratory tract infections range from 20 to 60 percent, and alcohol abuse has seen a occur among HIV-infected children, and strides made in sharp increase. In South Africa, suicide is found to be much reducing childhood mortality from lower respiratory tract more frequent among those who are HIV positive. infections during the 1980s and the early 1990s have been Neurological Disorders. The prevalence of epilepsy in Sub- reversed. In chapter 13, on tuberculosis, Dye and his Saharan Africa ranges from 2.2 to 58.0 per 1,000 people. colleagues discuss how people latently infected with Stroke has been found to be as common in Sub-Saharan Mycobacterium tuberculosis are at greater risk of developing Africa as in the West. The leading causes of neurological dis- active tuberculosis if their immune systems are also weak- orders are infections during pregnancy, neonatal infections, ened with HIV infection. Consequently, the tuberculosis and sequelae to the disorders that cause high under-five caseload has increased by a factor of five or more in the mortality. Challenges to the management of neurological countries of eastern and southern Africa most affected disorders include the lack of adequately trained personnel by HIV. able to recognize and manage the disorders, lack of equip- Malaria has a two-way relationship with HIV/AIDS. ment necessary to confirm a neurological diagnosis, and Anemia resulting from malaria increases the risk for HIV unavailability of the common drugs that would control infection through increased use of blood transfusions. In epilepsy. the review of malaria (chapter 14), Snow and Omumbo Violence and Injuries. Intentional injuries (violence) report an odds ratio for HIV infection of 3.5 for malaria resulted in the deaths of more than 300,000 people in Africa patients transfused once, 21.5 for those transfused twice, in 2000. Intentional injuries also are estimated to result in and 43.0 for those transfused three times during a single at least 6.2 million disabled or incapacitated people, 20 times admission. HIV infection, in turn, increases the risk of the number of deaths. Road traffic injuries, burns, drown- malaria, which is associated with higher density of para- ing, war, and homicide are the major causes of injury sitemia and more severe symptoms of malaria in adults. mortality in Sub-Saharan Africa. HIV/AIDS not only affects the incidence of communica- ble diseases but is also a risk factor for several noncommu- nicable diseases. As discussed in chapter 10, children with KEY DEVELOPMENTS SINCE DMSSA-1 HIV infection are at special risk for developmental disabili- ties. Low birthweight, prematurity, poverty, malnutrition, This section deals with the changes in the overall socioeco- and micronutrient deficiencies, more frequently seen in nomic environment that have had a major impact on preva- HIV-infected children, are likely to compromise early child lence of diseases in Sub-Saharan Africa, such as economic development. Maternal-child interaction is also affected; Changing Patterns of Disease and Mortality in Sub-Saharan Africa: An Overview | 3 even HIV-uninfected children of HIV-infected mothers are Figure 1.1 Real GDP per Capita Growth, by Region, 1991­2015 at higher risk for cognitive and language delays. 8 Kaposi's sarcoma, now ranked first for male cancers and third for female cancers in the region, is also associated with 6 HIV/AIDS. Prior to the epidemic, this was a rare cancer, but 4 it has increased twentyfold, and in countries with a high year prevalence of HIV, Kaposi's sarcoma is the leading cancer in per % 2 children. As discussed by Mbewu and Mbanya (chapter 21), 0 30 percent of those living with HIV show evidence of car- 2 diac involvement. Mental health also shows the impact of a an e e a and fic and a and and h Asi HIV/AIDS: psychiatric sequelae of HIV/AIDS include ries ries Asi ope -incom Paci erica bbean SaharAfrica e-incom Eur ral Asi i Sout low count Car e Eastth Africa Sub- count depression, anxiety disorders, manic symptoms, and atypi- East Cent Lat in Amthe MiddlNor middl cal psychosis. region Maternal HIV infection compromises the provision of 1991­2000 2001­04 2005­15 care and undermines global cognitive development even in the uninfected children. HIV-infected infants demonstrate Source: World Bank 2005a. lower mental and motor development (Baingana, Thomas, and Comblain 2005). Other effects of HIV on the nervous Figure 1.2 Real GDP per Capita, by Developing Region, system are discussed in chapter 23. 1980­2003 The direct impact of HIV on the incidence of and mor- 7,000 tality from both communicable and noncommunicable dis- eases is documented in the chapters that follow. HIV/AIDS 6,000 $ further affects health and mortality because of the social and 5,000 economic consequences of the disease, including a large 4,000 increase in the number of orphans, the burden on health international 3,000 services, the impact on human resources for health, and the 2,000 impoverishing consequences of the disease. The extraordi- constant 1,000 nary impact of HIV/AIDS has created a "development cri- sis" that extends far beyond its epidemiological effects. 0 1980 1985 1990 1995 2000 2003 year The Socioeconomic Context Sub-Saharan Africa other low-income region other mid-income region Growth in GDP per capita in low-income countries in Sub- Source: World Bank 2005c. Saharan Africa has continued to lag behind most other regions (figure 1.1), and real per capita GDP growth was negative for the period 1991 to 2000. Growth accelerated (table 1.1). Growth rates have also been more volatile: of during the first few years of the twenty-first century but still the 45 Sub-Saharan African countries, only 5 consistently lagged behind that of all other regions except Latin America recorded real per capita growth rates above 2 percent per and the Caribbean in 2004; the World Bank predicts that it year (Botswana, Cape Verde, Mauritius, the Seychelles, and will remain slow until 2015. In the 1980s, per capita income Swaziland), whereas nearly three-quarters of the countries expressed in purchasing power parity (PPPs, international experienced at least one year of per capita growth lower than dollars) was higher in Africa than in other low-income minus 10 percent (World Bank 2005a). countries, but it has gradually deteriorated (figure 1.2) and, Closely linked to the low level of economic growth is the as of 2004, was well below that of other low-income coun- lack of progress in reducing poverty. Although most of the tries (World Bank 2005c). world is on track to achieve the Millennium Development Although some countries experienced rapid growth, Goal (MDG) of a 50 percent reduction in the number of more countries showed declines in real per capita income people living below $1 per day, poverty has been on the (expressed in US$) during both the 1980s and 1990s increase in Sub-Saharan Africa: in 1990, 44.6 percent of the 4 | Florence K. Baingana and Eduard R. Bos Table 1.1 Gross National Income, per Capita, 1980, 1990, 2003 population lived below the $1 per day line; this had (current US$) increased to 46.4 percent by 2003. There is little doubt that slow economic growth and Percentage change increasing poverty are related to slow progress in health out- Country 1980 1990 2003 1980­90 1990­2003 comes. Wagstaff and Claeson (2004) summarized findings Angola -- 820 740 -- 10 on income, coverage of interventions related to health, and Benin 410 370 440 10 19 health outcomes, documenting that higher incomes lead to Botswana 1,120 2,750 3,530 146 28 improved access to and use of preventive and curative inter- Burkina Faso 290 330 300 14 9 Burundi 220 220 90 0 59 ventions, such as antenatal care, immunizations, use of Cameroon 620 950 630 53 34 treated bednets, and receipt of therapy for diarrhea and Cape Verde -- 980 1,440 -- 47 medicines for reducing fever. Income is also an important Central African Republic 340 470 260 38 45 determinant of access to nutritious food, which, in turn, Chad 240 270 240 13 11 leads to lower levels of malnutrition, a key risk factor for Comoros -- 550 450 -- 18 many childhood diseases. Congo, Dem. Rep. of 600 220 100 63 55 Congo, Rep. of 820 880 650 7 26 While some countries at lower-middle levels of income Côte d'Ivoire 1,140 780 660 32 15 have achieved good health outcomes, such examples are rare Equatorial Guinea -- 350 -- -- -- for the countries with the lowest incomes. A basic package Eritrea -- -- 190 -- -- of health interventions would in the case of the poorest low- Ethiopia -- 170 90 -- 47 income Sub-Saharan African countries overwhelm public Gabon 4,800 4,800 3,340 0 30 health budgets, and prospects for scaling up public health Gambia, The 380 310 270 18 13 Ghana 420 380 320 10 16 services from domestic resources are unfavorable. Guinea -- 460 430 -- 7 Guinea-Bissau 150 220 140 47 36 The Demographic Context Kenya 440 380 400 14 5 Lesotho 500 650 610 30 6 Sub-Saharan Africa is the "youngest" of the World Bank Liberia 530 -- 110 -- -- regions, as measured by the proportion of the population Madagascar 450 240 290 47 21 below age 15 and by the median age of the population. Malawi 190 200 160 5 20 Mali 270 270 290 0 7 About 44 percent of the population is younger than 15 Mauritania 450 540 400 20 26 (compared with 28 percent globally), and the median age of Mauritius -- 2,300 4,100 -- 78 the population is just 17.5 years (compared with 27 years Mozambique -- 170 210 -- 24 globally; figures 1.3, 1.4). In countries such as Uganda and Namibia -- 1,720 1,930 -- 12 Niger, the proportion below age 15 is close to 50 percent of Niger 440 310 200 30 35 the population. Fertility in Sub-Saharan Africa continues to Nigeria 780 270 350 65 30 Rwanda 250 370 220 48 41 be the highest in the world despite some decline in recent São Tomé and Principe -- 430 300 -- 30 years. From 1990 to 2003 the total fertility rate (TFR) Senegal 530 720 540 36 25 declined somewhat, but it is still higher now than in any Seychelles 2,080 5,020 7,490 141 49 other region in 1990 (figure 1.5). Sierra Leone 380 200 150 47 25 The youthfulness of the population reflects fertility and Somalia 100 130 -- 30 -- mortality rates, which in turn have an impact on the epi- South Africa 2,550 2,890 2,750 13 5 Sudan 470 570 460 21 19 demiological characteristics of the population. High fertili- Swaziland 960 1,190 1,350 24 13 ty and high adult mortality lead to a high proportion of Tanzania -- 190 300 -- 58 young people, who are much less likely to be vulnerable to Togo 450 440 310 2 30 chronic diseases that typically affect the adult and elderly Uganda -- 320 250 -- 22 populations. Epidemiology and demography thus interact Zambia 630 450 380 29 16 to generate the overall disease and mortality patterns in Zimbabwe 950 880 -- 7 -- All Sub-Saharan Africa 734 825 860 12 4 which infectious diseases are dominant over noncommuni- (unweighted) cable diseases and conditions. Population growth averaged 2.5 percent during 1990 Source: World Bank 2005c. Note: -- not available. and 2003 for the region as a whole, exceeding 3 percent in Changing Patterns of Disease and Mortality in Sub-Saharan Africa: An Overview | 5 Figure 1.3 Population below Age 15, 2003 Figure 1.5 Total Fertility Rate, 1990 and 2003 50 7 45 6 40 35 5 30 woman 4 population 25 per 3 total 20 of 15 births % 2 10 1 5 0 0 haran ca nd a Asi rld ribbean st Asi and and Asia and and and ca e and a wo ic income b-Sa Afri uth rica a andcific aharanfrica Asia So Pa l Asi A East South AsiaPacif Su Ea Europntra high income Sub-S h Africa Americaaribbean Europe high C East Mi ddleNo East aAfri rth Latin Ame the Ca Ce MiddleNort Central Latin the region region Source: United Nations 2005. TFR, 1990 TFR, 2003 Source: United Nations 2005. Figure 1.4 Median Age of Population, 1990 and 2003 40 Figure 1.6 Age Pyramid for Botswana, 2005, with and without AIDS 35 30 2005 25 80 age 20 70­74 Males Females 15 60­64 10 5 50­54 0 e d age 40­44 incom and ia ia As t and ca As ia andific ca and ca aharan worl high Afri Europe t As Pac eri bbean South e Eas Afri Central 30­34 Eas n Am Cari Lati the Middl Sub-S North region 20­24 median age, 1990 median age, 2003 10­14 Source: United Nations 2005. 0­4 140 120 100 80 60 40 20 0 20 40 60 80 100 120 140 number of people (thousands) countries such as Chad, the Republic of Congo, The without AIDS with AIDS Gambia, and Niger. At a rate of 2.5 percent, the population would double in less than 28 years. However, population Source: U.S. Census Bureau 2004. growth rates are projected to fall precipitously in countries in which HIV/AIDS has infected a large number of people. years, age structures of the affected countries will become World Bank projections for the region as a whole show the characterized by an unusually small number of adults, as population growth rate declining to 2.0 percent during shown in the age pyramid for Botswana (figure 1.6). 2000­10, and 1.9 percent during 2010­15. In the most affected countries in southern Africa, World Bank projec- Increasing International Attention to Health tions show a decline to between 0.2 and 0.5 percent growth in Sub-Saharan Africa per year. Other agencies that have published demographic projections show an even greater impact of AIDS mortality, In the years since the publication of DMSSA-1, the attention leading to population decline by 2010 in some countries. being paid to health conditions in Sub-Saharan Africa has Due to the high mortality of AIDS during the young adult rapidly increased, as evidenced by the number of studies 6 | Florence K. Baingana and Eduard R. Bos Box 1.2 The Health-Related Millennium Development Goals and Indicators · Goal 1: Eradicate extreme poverty and hunger · Goal 6: Combat HIV/AIDS, malaria, and other ­ Target is to cut in half the proportion of people diseases who suffer from hunger between 1990 and 2015. ­ Target is to have halted and begun to reverse the Progress is measured by the prevalence of under- spread of these diseases by 2015. weight children under five years of age. · Goal 7: Ensure environmental sustainability · Goal 4: Reduce child mortality ­ Target is to cut in half the proportion of people ­ Target is to reduce the under-five mortality rate without sustainable access to safe drinking water by two-thirds between 1990 and 2015. by 2015. · Goal 5: Improve maternal health · Goal 8: Develop a global partnership for development ­ Target is to reduce the maternal mortality ratio by ­ Target is to provide access to affordable essential three-quarters between 1990 and 2015. drugs in developing countries. and reports, new initiatives that draw attention to particular Africa. Traditional donors, such as bilateral development diseases, and increased financing from donor countries, agencies, the World Bank, and regional development banks, foundations, and multilateral agencies. have also increased financing for health, and the joint Many reports have either explicitly focused on Africa or WHO­World Bank High-Level Forum on the Health MDGs have focused on health conditions in poor countries, leading is considering new mechanisms to expand the availability of to a strong emphasis on Africa. Among the more prominent resources to combat communicable diseases. recent studies are the 2001 report Macroeconomics and An important influence on priorities for the global health Health: Investing in Health for Economic Development agenda are the MDGs, endorsed by 147 heads of state at the (Commission on Macroeconomics and Health 2001); the UN Millennium Summit of September 2000. The goals 2005 report Our Common Interest (Commission for Africa include numerical targets that are to be achieved between 2005); and World Bank studies and publications, such as 1990 and 2015. Of the eight goals, three are directly con- the 1998 publication Better Health in Africa: Experiences and cerned with mortality and morbidity, and six have been Lessons Learned, the 2005 report Improving Health, identified as "health related" (box 1.2). The focus of the Nutrition, and Population Outcomes in Sub-Saharan Africa: MDGs on achieving health outcomes has increased the The Role of the World Bank, and the Global Monitoring awareness of the lack of progress in Sub-Saharan Africa. Report 2005: Millennium Development Goals--From Other low- and middle-income regions show progress Consensus to Momentum (World Bank 1998, 2005b, 2005a, toward some of the MDGs (although current trends indi- respectively). cate that not a single World Bank region is making suffi- New initiatives and partnerships formed or strengthened cient progress to reach all of them). Sub-Saharan Africa is during recent years have similarly provided advocacy for not on track to achieve a single one of the targets. Halfway increased attention to diseases of the poor, generally with a through the period from 1990 to 2015, not a single Sub- focus on Africa. Among these are partnerships that focus on Saharan Africa country is on track for the under-five mor- neglected diseases that mostly affect Sub-Saharan Africa, tality rate target, and only one in four would achieve the such as guinea worm, trypanosomiasis, onchocerciasis, and malnutrition target on current trends. The increased focus schistosomiasis. Other global partnerships have increased on monitoring of trends has also provided evidence that the availability of pharmaceuticals at lower costs, through many countries in the region have worse indicators than pooled procurement, for diseases such as malaria and tuber- they did 15 years ago. culosis and for vaccine-preventable diseases. Foundations and funds, such as the Bill & Melinda Gates Foundation or Expanding Data Collection Efforts the Global Fund to Fight AIDS, Malaria, and Tuberculosis, have made large amounts of new financing available to Efforts to collect more data on health outcomes have inten- address diseases that disproportionately affect Sub-Saharan sified over the past decade, and as a result DMSSA-2 is more Changing Patterns of Disease and Mortality in Sub-Saharan Africa: An Overview | 7 empirically based than the previous edition. Household sur- Table 1.2 Overview of Health Worker Vacancy Rates for Four veys, including the Demographic and Health Surveys, the Countries UNICEF Mulitiple Indicator Cluster Surveys, the World Vacancy rates (percent) Bank's Living Standards Measurement Surveys, and other Health worker Ghana Lesotho Namibia Malawi surveys conducted by the World Health Organization as well as by country statistical offices, have vastly increased the Doctors 42.6 7.6 26.0 36.3 availability and quality of the data. Nurses 25.5 48.1 2.9 2.9 Demographic surveillance sites have joined in an alliance, Auxiliary nurses -- 5.4 0.6 18.4 called the INDEPTH Network, which has published stan- Doctor specialists 72.9 0.0 25.7 -- dardized reports on demographic indicators, including a set Other staff -- 29.9 25.3 62.8 of life tables. The network, which has grown to include 20 Source: Hongoro and McPake 2004. African sites, supports cross-site collaboration, capacity Note: -- not available. building, and dissemination of the collected data. Another area in which surveillance has greatly improved is HIV sur- veillance in antenatal clinics. Through annual reports of the the world disease burden, but only 1.3 percent of the share data, such surveillance has been used to document the sharp of the world's health workforce (Commission for Africa increases in HIV prevalence among pregnant women in 2005). Central to the problem are issues of supply, demand, southern African countries, as well as the decline in HIV and mobility (transnational, regional, and local). These prevalence in Uganda. Other areas of improvement over the include large differences in remuneration and nonreward- past decade include the surveillance and reporting of can- ing work in the low-income countries juxtaposed with a cers, from an increased number of cancer registries, and growing demand for skilled workers, in particular, nurses, in injuries, from injury surveillance systems. Advances have the high-income countries (Joint Learning Initiative 2004). also been achieved in malaria mapping and in the estima- The problem of low staff numbers is compounded by low tion of diabetes and lung disease incidence. morale and skills and the maldistribution of staff geograph- Nevertheless, the availability of morbidity and mortality ically. Further challenges are the wars and other internal data is far from sufficient for monitoring disease outbreaks, conflicts that adversely affect health infrastructure, services, the impact of health interventions, or even annual monitor- and personnel retention. The HIV epidemic increases the ing of incidence and prevalence of most diseases. Routine workload, and AIDS mortality has reduced the number of vital registration is still absent in almost all countries (except health workers. In countries such as Malawi and Zambia, it Mauritius and the Seychelles), although progress has been is estimated that the illness of health workers has increased made in mortality registration in South Africa. One conse- five- to sixfold (Padarath et al. 2003). quence of this lack is the general unavailability or reliability of the denominators needed to estimate overall mortality Conflicts, Refugees, and Internally Displaced People or cause-specific rates. Efforts to expand the coverage of vital registration beyond urban areas would have substantial In the years since DMSSA-1 was published, the continent payoffs for improving the quality of epidemiological has undergone numerous armed conflicts, including civil information. wars and genocide. Since 1980, more than 30 wars have plagued Africa. It is estimated that as of the end of 2003, 16 million people in Sub-Saharan Africa had been displaced Human Resources for Health: A Worsening Crisis? through conflict (WHO 2002). Low-income countries are Human resources have been described as "the heart of the disproportionately affected by conflicts. Fifteen countries in health system in any country," and "the most important the region had a major conflict between 1990 and 2003 aspect of health care systems" (Hongoro and McPake 2004). (UNICEF 2005). Table 1.3 illustrates the relative global The recent study Human Resources for Health: Overcoming burden of conflict-related deaths by region. the Crisis, by the Joint Learning Initiative (2004), suggests Injuries due to collective violence are concentrated in that both the number and the skill levels of health workers Sub-Saharan Africa. In the last decade, the bulk of lives lost in Sub-Saharan Africa are far below what is needed to to war injuries in Africa have resulted from conflicts in the reduce mortality (table 1.2). The region has 25 percent of Democratic Republic of Congo, Liberia, and Rwanda. The 8 | Florence K. Baingana and Eduard R. Bos Table 1.3 Conflict-Related Deaths by Region the Central African Republic fell from 82 percent to 29 per- (per 100,000 people) cent, and in the Democratic Republic of Congo, from 79 percent to 33 percent (see chapter 12). The probability of Region No. of deaths surviving from age 15 to age 60 in 2000 was less than 50 per- High-income countries 0.0 cent in almost half of the Sub-Saharan Africa countries, due Low-income countries 6.2 in part to the conflicts. WHO Africa region 32.0 WHO Eastern Mediterranean region 8.1 WHO European region 4.2 REFERENCES Source: WHO 2002. Baingana, F., R. Thomas, and C. Comblain. 2005. HIV/AIDS and Mental legacy of war in the form of landmines continues to con- Health. Health Nutrition and Population electronic discussion paper. http://www.worldbank.org. tribute to mortality in the continent. As of October 2004, Commission for Africa. 2005. Our Common Interest: Report of the Com- 1.2 million Sudanese had been uprooted from their homes, mission for Africa. London: Commission for Africa. http://www. many killed by militias, and those who found their way into commissionforafrica.org. Chad faced disease, poor nutrition, and inadequate shelter. Commission on Macroeconomics and Health. 2001. Macroeconomics and Health: Investing in Health for Economic Development. Geneva: WHO. In a typical five-year war, the under-five mortality increases Feachem, R. G., and D. T. Jamison. 1991. Disease and Mortality in Sub- by 13 percent and adult mortality even more. During the Saharan Africa. Washington, DC: World Bank. first five years of peace, the average under-five mortality was Hongoro, C., and B. McPake. 2004. "How to Bridge the Gap in Human found to be 11 percent higher than the corresponding level Resources for Health." Lancet 364: 29­34. before the war. Sexual violence during conflicts increases the Joint Learning Initiative. 2004. Human Resources for Health: Overcoming the Crisis. Cambridge, MA: Harvard University Press. spread of HIV (UNICEF 2005). Padarath A., C. Chamberlain, D. McCoy, A. Ntuli, M. Rowson, and R. In Sub-Saharan Africa, for children who survive the first Loewenson. 2003. "Health Personnel in Southern Africa: Confronting four years of life, injury becomes the most likely cause of Maldistribution and Brain Drain." Discussion paper 3, Equinet Africa, Training and Research Support Centre (TARSC), Harare, Zimbabwe. disability and death. Most intentional injuries are caused by http://www.equinetafrica.org/bibl/resources.php. war; it is estimated that 120,000 to 200,000 child soldiers age UNICEF (United Nations Children's Fund). 2005. The State of the World's 5 to 16 years are participating in conflicts, putting them at Children: Childhood under Threat. New York: UNICEF. risk for bullet and shrapnel wounds, burns, and land mine United Nations. 2005. World Population Prospects: The 2004 Revision. New York: United Nations. injuries (UNICEF 2005). Psychosocial and mental disorders U.S. Census Bureau. 2004. International Programs Center, AIDS surveil- resulting from conflicts had affected 15.5 percent of the lance database. http://www.census.gov/ipc/www/hivaidsn.html. population in Rwanda five years after the genocide; depres- Wagstaff, A., and M. Claeson. 2004. The Millennium Development Goals for sion, anxiety, and PTSD can range from 20 to 60 percent in Health: Rising to the Challenges. Washington, DC: World Bank. conflict-affected populations (Baingana, Thomas, and World Bank. 1998. Better Health in Africa: Experiences and Lessons Learned. Washington, DC: World Bank. Comblain 2005). ------. 2005a. Global Monitoring Report 2005: Millennium Development The most dramatic outbreak of a diarrhea epidemic Goals--From Consensus to Momentum. Washington, DC: World Bank. occurred in July 1994 among Rwandan refugees in Goma, ------. 2005b. Improving Health, Nutrition and Population Outcomes in Democratic Republic of Congo, when almost 50,000 Sub-Saharan Africa. Washington, DC: World Bank. refugees died (see chapter 9). Conflicts have also had an ------. 2005c. World Development Indicators 2005. Washington, DC: World Bank. http://devdata.worldbank/dataonline/. impact on immunization rates. From 1990 to 2000 the vac- WHO (World Health Organization). 2002. World Report on Violence and cination rates for diptheria, pertussis, and tetanus (DPT) in Health. Geneva: WHO. Changing Patterns of Disease and Mortality in Sub-Saharan Africa: An Overview | 9 Chapter 2 Levels and Trends in Mortality in Sub-Saharan Africa: An Overview Jacob Adetunji and Eduard R. Bos One of the major achievements of the twentieth century in whereas others are associated with increasing exposure to Sub-Saharan Africa is the unprecedented decline in mortal- risk factors that lead to increased morbidity and mortality ity and the corresponding increase in the expectation of life (such as increasing exposure to risks for noncommunicable at birth. At the dawn of the twentieth century, Sub-Saharan diseases or the spread of new and reemerging communica- Africa was characterized by extremely high under-five mor- ble diseases). Therefore, monitoring mortality levels and tality levels and by low life expectancy at birth. By the end of trends in the Sub-Saharan region provides not only a direct the century, however, mortality among children under five reflection of the health status of populations but also an had decreased from about 500 per 1,000 live births to about indirect gauge of the effects of economic, political, and epi- 150 (World Bank 2005). Similarly, the average length of life, demiological turbulence that faced the region. which was less than 30 years about 100 years ago, had increased to more than 50 years by the early 1990s. Much of the mortality decline happened in the second half of the INDICATORS OF MORTALITY LEVELS AND TRENDS twentieth century, the fastest rate of decline occurring in the first decades after World War II (Hill 1991). In the 1990s, In this overview chapter, two indicators of mortality are mortality decline stalled for the region overall, with many used to assess levels and trends for Sub-Saharan Africa, its countries experiencing reversals in the upward trend in life subregions, and countries. The infant mortality rate, calcu- expectancy largely because of AIDS mortality. lated as the proportion of newborns in a given period that This overview focuses on the period between 1960 and do not survive to their first birthday, is a standard measure 2005. This period roughly corresponds to the postcolonial not affected by age structure and therefore suitable to use for era in many countries in the region, in which large economic comparisons over time and across regions. Life expectancy and social changes occurred. Some of these changes were at birth, calculated as the average number of years a beneficial to the health of the population (such as economic newborn would live if subject to the mortality rates for a growth and increasing access to health interventions), given year, is used to compare the force of mortality across 11 the entire age spectrum. The dearth of reliable data is one of Table 2.1 Life Expectancy at Birth for World and UN Regions, the main problems confronting the study of mortality levels 1960­2005 and trends in Sub-Saharan Africa. Although vital registra- Region 1960­69 1970­79 1980­89 1990­99 2000­04 tion systems exist in most countries in the region, they usu- World 52.5 58.1 61.4 63.7 65.4 ally do not produce reliable data. In the absence of reliable Sub-Saharan Africa 42.4 46.3 49.0 47.6 45.9 vital registration systems and good quality census data that Asia 48.5 56.4 60.4 64.0 67.3 are needed for direct calculation of infant and child mortal- Europe 69.6 71.0 72.0 72.6 73.7 ity rates, demographers have developed indirect methods Latin America 56.8 60.9 64.9 68.3 71.5 for obtaining these vital statistics from incomplete and often and Caribbean defective data. However, over the past 30 years, information Northern America 70.1 71.6 74.3 75.5 77.6 available for the study of mortality patterns, particularly Oceania 63.7 65.8 69.3 71.5 74.0 among children under age five, has improved dramatically. Source: United Nations 2005. The improvement in information is largely due to the implementation of large-scale household survey programs, such as the World Fertility Surveys (WFS) program of 1972­84, the Demographic and Health Surveys (DHS), and Africa and other regions, followed by a comparison of these UNICEF's Multiple Indicator Cluster Surveys (MICS). Of all mortality indicators for subregions within Sub-Saharan these survey programs, the DHS has had the largest impact Africa. on data availability, analysis, and report dissemination. About 70 DHS surveys have been conducted in 33 of the 46 major countries in Sub-Saharan Africa. Sub-Saharan Africa Relative to Other, Less Developed Apart from the DHS-type surveys, Sub-Saharan Africa Regions has an extensive network of longitudinal study sites. At least Sub-Saharan Africa is, by far, the region of the world with 19 such study sites exist in the region and their data have the highest level of mortality. Overall life expectancy at birth been invaluable in deriving mortality estimates by age as well is 46 years, whereas in Asia, the region with the second as model life tables that show how the age pattern of African lowest life expectancy, it is 67. mortality differs from the model life tables constructed by As shown in table 2.1, the disparity between Sub-Saharan Coale and Demeny (1983) and United Nations model life Africa and other regions of the world has widened since the tables (INDEPTH Network 2001). The main problem with 1960s. In that decade the difference in life expectancy with this source is that most of these longitudinal study sites are the Asian region was only 6 years, but this has grown to based in rural settings and are scattered throughout the almost 21 years now. And, whereas all other regions have whole region and therefore provide estimates of unknown experienced uninterrupted increases in life expectancy, in generalizability. The locations of the sites are neither system- Sub-Saharan Africa life expectancy peaked in the early 1990s atically planned to represent the Sub-Saharan Africa region at 50 years, and has since fallen back by almost 4 years. nor do they adequately represent the countries in which they Declines in infant mortality rates in Sub-Saharan Africa are located. started to slow down considerably in the 1990s. These slow In this chapter the estimates for countries and subre- declines have meant that Sub-Saharan Africa has lagged gions are those issued most recently by the United Nations more and more behind other regions and hence the mortal- Population Division; the estimates are based on a variety of ity gap has widened (table 2.2). sources, including surveys, censuses, and demographic modeling. The delineation of geographic subregions used are those defined by the United Nations. Subregional Differences in Mortality In Sub-Saharan Africa as a whole, infant mortality rates MORTALITY LEVELS AND TRENDS declined from 149 per 1,000 live births in the 1960s to about 101 in 2005--a 32 percent decline over a period of 35 years. The following section will provide a comparison of indica- Toward the end of the last decade of the twentieth century, tors of mortality trends discussed above, first comparing the decline in infant mortality rates leveled off, decreasing trends in life expectancy and infant mortality in Sub-Saharan only slightly for the region as a whole. 12 | Jacob Adetunji and Eduard R. Bos Table 2.2 Infant Mortality Rates for World and UN Regions, Table 2.4 Life Expectancy at Birth for Sub-Saharan Africa 1960­2005 and UN Subregions, 1960­2005 (per 1,000 live births) Region, subregion 1960­69 1970­79 1980­89 1990­99 2000­04 Region 1960­69 1970­79 1980­89 1990­99 2000­04 Sub-Saharan Africa 42.4 46.3 49.0 47.6 45.9 World 119 93 78 66 57 Eastern Africa 43.4 47.3 49.4 46.7 45.7 Sub-Saharan Africa 149 130 115 107 101 Middle Africa 41.0 45.3 47.0 44.3 43.4 Asia 123 96 77 63 54 Southern Africa 50.7 54.4 59.6 59.6 47.7 Europe 33 23 17 11 9 Western Africa 40.3 43.9 47.1 47.2 46.3 Latin America 96 75 52 35 26 Source: United Nations 2005. and Caribbean Northern America 24 16 9 7 7 Oceania 49 43 36 32 29 Country Differences in Mortality Source: United Nations 2005. Figure 2.1 illustrates the differences in the levels and trends in the infant mortality rate in selected Sub-Saharan Africa countries. The rates vary from 15 in Mauritius, to 165 in Table 2.3 Infant Mortality Rates for Sub-Saharan Africa and UN Subregions, 1960­2005 Sierra Leone, and the rates of change from 1960 to the pres- (per 1,000 live births) ent differ from about 20 percent in the Democratic Republic of Congo, Liberia, Rwanda, and Sierra Leone to over 50 per- Region, subregion 1960­69 1970­79 1980­89 1990­99 2000­04 cent in countries in Southern Africa. It is noteworthy, Sub-Saharan Africa 149 130 115 107 101 however, that infant mortality has declined in all countries Eastern Africa 144 124 112 101 93 since 1960. Middle Africa 156 131 121 122 116 Figure 2.2 shows country patterns in life expectancy at Southern Africa 90 78 58 47 45 birth. The range in current levels is about 35 years, from a Western Africa 165 145 128 119 114 high of 72 in Mauritius to a low of 37 in Zimbabwe and Source: United Nations 2005. Zambia. Recent trends are clearly negative in many coun- tries, where increases in adult mortality resulting from AIDS have led to a decline in overall life expectancy. Most of these In regard to subregional disparities, infant mortality rates countries experienced the highest life expectancies during are highest in West Africa and in Middle Africa and have 1985 to 1990 and have since declined to below the levels consistently been so from 1960 (table 2.3). The infant mor- in 1960. tality rate declined somewhat faster in West Africa, and as a result, Middle Africa is currently the subregion with the Figure 2.1 Infant Mortality Rate in Selected Countries, highest rate. Of all subregions of Sub-Saharan Africa, coun- 1960­2005 tries in Southern Africa have had the lowest infant mortality (per thousand) rates. For example, in 1960 the rate was 42 percent lower 300 than in other subregions, and even with increasing overall 250 mortality in the 1990s, the infant mortality rate in Southern Africa was still less than half the average for Sub-Saharan rate 200 Africa in 2000. 150 Life expectancy at birth has increased 3.5 years for the mortality continent as a whole since 1960, but it is now lower in 100 infant Southern Africa than in the 1960s (table 2.4). All the subre- 50 gions reached peak levels of life expectancy about 1990, but 0 they have since shown a decline, largely due to AIDS mor- 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 year tality. Nowhere has the decrease in life expectancy been steeper and greater than in Southern Africa, where 40 years Mauritius Namibia Tanzania of increases in life expectancy were reversed in a period of Benin Mali Sierra Leone 10 years. Source: Adapted from United Nations 2005. Levels and Trends in Mortality in Sub-Saharan Africa: An Overview | 13 Figure 2.2 Life Expectancy in Selected Countries REFERENCES 80 Coale, A., and P. Demeny. 1983. Regional Model Life Tables and Stable Populations. New York: Academic Press. 70 Hill, A. 1991. "Infant and Child Mortality: Levels, Trends and Data 60 Deficiencies." In Disease and Mortality in Sub-Saharan Africa, ed. R. G. Feachem and D. T. Jamison, 37­74. New York: Oxford University Press. 50 INDEPTH Network. 2001. Population and Health in Developing Countries. Vol. 1. Ottawa: International Development Research Centre. 40 United Nations. 2005. World Population Prospects. The 2004 Revision. New expectancy York: United Nations. life 30 World Bank. 2005. World Development Indicators. Washington, DC: World 20 Bank. 10 0 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 year Mauritius Namibia Tanzania Benin Mali Sierra Leone Source: Adapted from United Nations 2005. 14 | Jacob Adetunji and Eduard R. Bos Chapter 3 Trends in Child Mortality, 1960 to 2000 Kenneth Hill and Agbessi Amouzou Under-five mortality, the probability of dying between birth Sub-Saharan Africa. However, the quantity, timeliness, and and age five expressed per 1,000 live births, and infant mor- quality of available information vary widely by country. tality, the probability of dying before age one expressed per Figure 3.1 contrasts the data availability for the Republic of 1,000 live births, are widely used as measures of children's, Congo and Kenya. The only information available for and more broadly a population's, well-being. Reduction the Republic of Congo is a set of indirect estimates derived of the under-five mortality rate (U5MR) by two-thirds from data collected by the 1974 census, effectively covering between 1990 and 2015, equivalent to an annual average rate only the period 1960 to 1970. Kenya, in contrast, has esti- of reduction of 4.3 percent, is one of the six health-related mates from several censuses, a World Fertility Survey Millennium Development Goals (MDGs). Data indicate (WFS), a National Demographic Survey (NDS), and three that some 11 million children under the age of five die Demographic and Health Surveys (DHSs). For the Republic annually in the world as a whole, and more than 10 million of Congo, there exists no possibility of consistency checks, of these deaths occur in the developing world. Sub-Saharan and there has been no empirical basis for estimating child Africa is the region most affected and accounts for more mortality since about 1970. For Kenya, however, the differ- than one-third of deaths of children under the age of five ent data sources provide a large number of estimates, not all (Hill et al. 1999). Some two-thirds of the child deaths in the of which are mutually consistent, for overlapping time peri- developing world are caused by diseases (predominantly ods and a considerable density of observations covering the acute respiratory infections, diarrhea, and malaria) for early 1960s to the late 1990s. which practical, low-cost interventions, including immu- The multiplicity and in some cases inconsistency of nization, oral rehydration therapy (ORT), and antibiotics, U5MR estimates from different sources has made the deter- exist (Jones et al. 2003). mination of national trends problematic. Hill and Yazbeck The quality and quantity of data on child mortality have (1994), and subsequently Hill and colleagues (1999), devel- increased dramatically over the last 30 years, particularly in oped and applied an explicit, objective, and replicable 15 Figure 3.1 Contrasting Data Availability: The Republic of year to births in the same year obtained from civil registra- Congo and Kenya tion data. The mortality of children after infancy is typically (per 1,000 births) obtained from civil registration information on deaths of a. The Republic of Congo young children by age and from population census infor- mation on the size of the population of those ages exposed 250 to the risk of dying.1 Thus, civil registration data provide all 5 200 the information needed to measure infant mortality, which age by can therefore be readily calculated annually, but measure- 150 ment of mortality after infancy requires additional informa- dying of tion on population sizes. 100 In countries where the registration of vital events is not complete, the registration of infant deaths is often less com- probability 50 plete than the registration of births, with the result that the 0 1960 1970 1980 1990 2000 registered IMR underestimates the true value. In these year countries, estimates of infant and under-five mortality are Census 74 Fitted line typically obtained instead from one or more of three types of survey data. Most similar to registration data is that b. Kenya obtained from the longitudinal or prospective sample 300 survey. A sample of the national population is followed over a period of time, with all vital events being recorded. Such 5 250 data provide the basis for calculating the conventional IMR age by 200 as the ratio of infant deaths to births, and also provide the dying basis for calculating mortality rates after infancy, since 150 of population numbers are also available. Such surveys have 100 not been widely used, partly because they are expensive to probability 50 mount, and partly because they require careful supervision over an extended period to provide good data on trends. 0 1960 1970 1980 1990 2000 The second data source is a retrospective sample survey year that collects full birth histories. Each mother is asked for Census WFS/DHS direct WFS/DHS indirect information on the date of birth and, if relevant, the age at NDS indirect Fitted line death of every live-born child she has had. In the 1970s and Source: Updated version of database compiled by Hill and colleagues (1999). early 1980s the WFS program, and more recently the DHS project, collected such data in many developing countries. Both IMRs and U5MRs for periods up to 15 years before the methodology to derive a single consistent time series of esti- survey can be calculated from the data, dividing deaths for mates for infant and under-five mortality from the assem- given ages and time periods by exposure to risk (expressed bled data. as person-years of life lived) of the reported children In this chapter, the estimates by Hill and colleagues (Somoza 1980). However, the collection of such informa- (1999) for countries of Sub-Saharan African are updated. tion by surveys is complex and requires high levels of inter- Country data that have become available since that study are viewer quality and training. The surveys are therefore quite added in order to provide more complete information expensive and can cover only relatively small samples. about levels and trends of under-five mortality since 1990. The third data source is a retrospective survey that col- lects summary birth histories. Each woman surveyed is DATA SOURCES AND METHODS asked for very simple information: her age, the total number of children she has borne, and the number of those children In countries with accurate registers of births and deaths, that have died; in short, a summary birth history with no infant mortality year by year is measured by the infant mor- information about individual children. For a particular age tality rate (IMR), the ratio of deaths of infants under one group of women, the proportion of children dead depends 16 | Kenneth Hill and Agbessi Amouzou primarily on two things: the level of under-five mortality, of estimation of child mortality from information on chil- and the distribution of the children by how long they had dren ever born and children surviving (see United Nations been exposed to risk. William Brass (1964) first developed a 1983). The MICS surveys are also nationally representative method for deriving estimates of child mortality from such surveys, conducted by UNICEF in many countries in Sub- data. The Brass method and developments of it (Sullivan Saharan Africa. Most MICS surveys record only summary 1972; Trussell 1975; United Nations 1983) adjust the pro- birth histories for each woman of reproductive age, thus portions of children dead by age group of mother for an allowing only indirect estimation of infant and child mor- estimated exposure distribution in order to arrive at pure tality. Only one country in the region--South Africa--is measures of under-five mortality and of reference dates for thought to have vital registration data that are complete these measures. The adjustment process assumes certain enough to provide valid estimates of the IMR (Rob patterns of fertility and under-five mortality by age, and Dorrington, personal communication 2004). These esti- results can be quite sensitive to the choices made. The mates are not used here, since there is no satisfactory way of methodology is better suited to estimating under-five mor- knowing at what point in the recent past the infant mortal- tality than infant mortality (Hill 1991). ity information became complete. Hill and colleagues (1999) reviewed more completely Nine of the Sub-Saharan African countries that Hill and methods for measuring under-five mortality in countries colleagues (1999) included in their study have no informa- lacking accurate vital registration data, and Hill (1991) tion on child mortality more recent than that already reviewed their various advantages and disadvantages. analyzed. Seven countries--Angola, Comoros, Gabon, Guinea-Bissau, Somalia, São Tomé and Principe, and South Africa--that they did not include are in the present analysis. Data Used The database they used has been updated with direct or The basic data used in this chapter are observations of indirect estimates of infant and child mortality based on U5MR derived from full or summary birth histories. A recent data for 27 countries. Some small countries--Cape summary birth history provides seven observations of Verde, Djibouti, Equatorial Guinea, Mauritius, Réunion, U5MR, one for each five-year age group of mothers and the Seychelles--are not included because they did not between ages 15 and 49. A full birth history typically pro- have appropriate data or were not regarded as being repre- vides three observations of U5MR, for the periods 0 to 4, 5 sentative of Sub-Saharan Africa. In sum, a total of 153 data to 9, and 10 to 14 years before the survey, although obser- sets covering 42 countries of Sub-Saharan Africa are included vations are sometimes also calculated for the periods 15 to in this analysis. 19 and 20 to 24 years before the survey. The observations used are the same as those used by Hill and colleagues Methodology for Estimation of Trends (1999), updated with observations that have become avail- able more recently. The methodology adopted in this chapter is that used by Observations for time periods since the early 1990s come Hill and colleagues (1999; see appendix 3A for a detailed primarily from DHS surveys conducted since 1995 and from description). The chapter focuses primarily on under-five Multiple Indicator Cluster Surveys (MICS) conducted in mortality; all the observations used come from full or sum- 2000. A DHS is a nationally representative survey; such sur- mary birth history data collected by censuses or surveys, and veys have been conducted in many developing countries, it is argued that such methods of data collection provide and an increasing number of countries have conducted better estimates of U5MR than of IMR. For each country, more than one survey. A typical DHS survey records full many observations with a range of time reference dates are birth histories of women of reproductive age (15 to 49) available. These observations are not typically mutually con- along with the survival status of each child ever born to sistent, so some approach is needed to obtain a set of esti- them. Thus, DHS data provide a basis for both direct and mates from the various observations. In our approach, a indirect estimation of child and infant mortality. The survey regression model is fitted to the observations of U5MR by reports tabulate direct estimates for at least the three five- reference date. The model assumes that at the country level year periods preceding the survey. In this chapter we have the rate of change of U5MR is constant over defined time also used the birth history data as a basis for indirect esti- periods, the time periods themselves being determined by mation using the Manual X version of the Brass technique the number of observations available. Thus, a country with Trends in Child Mortality, 1960 to 2000 | 17 a large number of observations can have a more flexible where r 5 n 5 is the rate of change between year n 5 and time trend of U5MR than a country with few observations. year n, and [n] represents a year ending in 0 or 5. Estimates of U5MR are then derived from the model for Note that in this section, the unit of analysis is the coun- specified time points from 1960 to 2000. It is these estimates try, not the child. A country will appear in the analysis for a that are described below. The appendix shows country particular time point only if the U5MR is supported by examples of observations and the time trends fitted to them. an observation within two years of that time point. Thus country­time point estimates are selected for data availabil- ity, so that, for example, the countries contributing to RESULTS estimates for 1990 will not be the same as the countries con- tributing to estimates for 1995. Any association between lev- The U5MR is widely recognized as an important indicator els and rates of change of U5MR on the one hand and data of development. In high-mortality settings, it is preferable to availability on the other may cause a bias. the IMR as a single index because the U5MR captures the substantial mortality risk after 12 months of age (Hill 1995). Consequently, this section focuses on changes in U5MR Estimated Level of U5MR and IMR in 2000 derived from the country-specific analyses undertaken in Estimated levels of U5MR and IMR in 2000 are shown in this chapter, rather than on changes in the IMR. However, figure 3.2. The U5MR ranges from 47 per 1,000 births (South for completeness we do present country estimates of the lev- Africa) to 314 per 1,000 (Sierra Leone); the median is 153 per els of infant mortality in 2000, but it must be emphasized 1,000. For the median country, among 100 children born, that these estimates are derived from the U5MR estimates using the Coale-Demeny model life table system (Coale and Figure 3.2 Estimated Levels of Under-Five and Infant Demeny 1983). Tables 3B.1 and 3B.2 in appendix 3B show Mortality, by Region of Africa, 2000 country-specific estimates of U5MR and IMR, respectively, (per 1,000 births) for five-year time points from 1960 to 2000. Table 3B.1 also a. Under-five mortality shows the number of censuses or surveys for each country on which the estimates are based, and gives a subjective 350 assessment of the quality of the estimates, based on their 300 consistency. Figure 3B.1 provides graphs of available esti- 250 mates plotted against time for one example country in each 200 mortality category. 150 Levels and rates of change in U5MR over the 40 years from 1960 to 2000, estimated using the methodology under-five 100 described above, are summarized here in the form of box 50 plots2 for major regions of Sub-Saharan Africa, and for 0 Middle Eastern Southern Western total Sub-Saharan Africa as a whole. The United Nations (UN) region definitions of regions--Middle, Eastern, Southern, and Western--are used (see tables 3B.1 and 3B.2 for countries b. Infant mortality included in each region). 350 The model in equation 3.1 is used to estimate U5MR lev- 300 els for the mid-point of years ending in 0 or 5. 250 ln(5q0)i b0 b1(date)i b2(postk1)i 200 b3(postk2)i b4(postk3)i . . . mortality i (3.1) 150 infant Change is then measured between estimated values over 100 five-year periods, 1960­65, 1965­70, and so on. Rates of 50 change are calculated as 0 Middle Eastern Southern Western total 1 q [n] region 5 0 5 n r 5 ln (3.2) 5 5 0 q [n 5] Source: Tables 3B.1 and 3B.2. 18 | Kenneth Hill and Agbessi Amouzou more than 15 die before their fifth birthday. The comparable a whole, U5MR declined from 1960 to 2000. However, most figure for the developing world as a whole is less than 100 of the decline took place between 1965 and 1990, during per 1,000, and for the more developed world, less than 10 per which period the median U5MR dropped from 232 to 165 1,000. The interquartile range is from 124 to 217, indicating per 1,000. Since 1990 the downward trend has stalled. This continuing high child mortality in the majority of countries overall trend also characterizes each region, although at dif- covered. Infant deaths account for almost two-thirds of ferent levels and speeds. The countries of the Western region under-five deaths. The median IMR is 101 per 1,000 with an had the highest U5MR in 1960, with a median value of interquartile range from 80 to 133 per 1,000. about 291 per 1,000 live births. This level fell below 183 per There is considerable variability in the level of U5MR and 1,000 by 1990. The Middle region had a median rate of 268 IMR by region. The Middle and Western Africa regions have per 1,000 in 1960 and experienced a decline to 164 per 1,000 the highest median level of U5MR at 174 per 1,000 live in 1990. The median in the Eastern region oscillated around births, followed by the Eastern region with a median of 168 200 per 1,000 prior to 1975 before declining to 170 per 1,000 per 1,000. The Southern region exhibits a comparatively low in 1990. The Southern region had the lowest median U5MR rate of U5MR of 58 per 1,000. The Western region has the in 1960 (about 200 per 1,000) and experienced the sharpest highest level of IMR (a median of 117 per 1,000), followed decline, to about 70 per 1,000 by 1990. The Western and by the Middle and the Eastern, which have similar rates Southern regions thus experienced the fastest declines from (medians of 105 per 1,000). The Southern region has the 1960 to 1990, with the countries of the Middle and Eastern lowest level of IMR (a median of 46 per 1,000). The Western regions showing the slowest improvement. Declines appear region has the highest variability between countries in to have stalled in all regions in the 1990s. infant and under-five mortality, and the Southern region Figure 3.4 summarizes the annual rates of change in has the lowest. U5MR over five-year periods. For Sub-Saharan Africa as a whole, rates of decline averaged about 1 percent in the 1960s, increasing to close to 2 percent between 1970 and Trends in U5MR 1985, dropping back to about 1 percent between 1985 and Figure 3.3 presents levels of U5MR at time points separated 1990, and averaging less than 1 percent during the 1990s. by five years, from 1960 to 2000. For Sub-Saharan Africa as These rates of decline are much below the average annual Figure 3.3 Trends in Under-Five Mortality, by Five-Year Period and Region, 1960­2000 (per 1,000 births) 400 1960 1960 1965 1970 1965 350 1975 1975 1960 1965 1970 1980 2000 2000 300 1975 1985 1995 1995 1990 1980 1985 1990 1965 1970 1985 1990 1980 1995 2000 1960 250 1970 1985 1990 1975 2000 1980 1995 1960 1965 mortality 200 1970 -five 1975 under 150 1980 1985 1990 1995 2000 100 50 0 Middle Eastern Southern Western total region Source: Table 3B.1. Note: See endnote 2 for a description of a box plot. Trends in Child Mortality, 1960 to 2000 | 19 Figure 3.4 Annual Rate of Change in Under-Five Mortality, by Five-Year Period, 1965­2000 1995 5 1990 4 1990 1995 3 1995 1990 1995 1980 1985 1985 2 1980 U5MR 1990 1 1985 1990 1995 of 1985 1975 1980 1975 0 1965 1965 1965 1970 1970 1975 1965 1980 1970 change 1 1970 1975 1965 1970 1985 2 annual 1980 % 3 1975 4 5 6 Middle Eastern Southern Western total region Source: Table 3B.1. Note: See endnote 2 for a description of a box plot. decline of 4.3 percent required between 1990 and 2015 to but also about the age pattern of that mortality. The birth reach MDG targets. Indeed, between 1960 and 2000, the history data can be used to calculate the probability of dying median rate of decline in U5MR approached the required in the first year of life (the IMR) and the probability of rate of decline in only one region (Southern) for only one dying between exact ages one and five (the child mortality five-year period. Also, these rates of decline average only rate, or CHMR). Long data series for the developed world about half those of the developing world as a whole. The rel- indicate that as under-five mortality declines, CHMR atively poor performance of Sub-Saharan Africa in the tends to decline faster than IMR, leading to an increased 1990s is underlined by the observation that a large number concentration of mortality risk in the first year of life of countries, especially in the Middle, Eastern, and Southern (Hill 1991). The historical experience of the developed regions, have a positive rate of change in the 1990s, indicat- world also shows substantial variation in the relationship ing increasing U5MRs. In the Middle Africa region, the esti- between IMR and CHMR between populations (Coale and mates indicate increases in U5MR for Cameroon and Chad Demeny 1983). and a stall in the Democratic Republic of Congo (see the Figure 3.5 plots recorded CHMR against recorded IMR country-specific estimates in table 3B.1). In the Eastern for each DHS survey included in this analysis for three Africa region, the rise or stall is observed in Burundi, Kenya, Rwanda, Zambia, and Zimbabwe. In the Southern region, Figure 3.5 Relationships between CHMR and IMR, by Country Botswana and Swaziland are adversely affected, whereas in (per 1,000 live births) the Western region, Burkina Faso, Côte d'Ivoire, Mauritania, and Sierra Leone show increases. All these countries have 250 one or more things in common: they are seriously affected 200 by the human immunodeficiency virus and acquired immune deficiency syndrome (HIV/AIDS) epidemic or 150 have suffered an economic crisis or political instability that CHMR 100 created a civil war. It is likely that these factors have played a major role in the poor health performance observed in these 50 countries and in Sub-Saharan Africa as a whole. 0 50 100 150 200 Age Patterns of Under-Five Mortality IMR Eastern, Middle, and Southern Western Coast Western Sahel Birth histories, such as those collected by the DHSs, provide information not only about childhood mortality levels Source: Updated version of database compiled by Hill and colleagues (1999). 20 | Kenneth Hill and Agbessi Amouzou five-year periods prior to the survey, that is, for periods 0 As can be seen, the broad pattern of the relationship to 4, 5 to 9, and 10 to 14 years before data collection. between CHMR and IMR is similar for populations of Inspection of an early version of this figure using different Eastern, Middle, and Southern Africa on the one hand and symbols for observations from countries for five different the populations of Sahelian West Africa on the other, except regions of Sub-Saharan Africa (Eastern, Southern, Middle, that for a given level of IMR the CHMR in the former group Western Coast, and the western Sahel, using UN definitions of countries is about 30 per 1,000 lower than in the latter of regions except for the additional breakdown of Western group. The Western Coast, however, appears to have a very Africa into coastal and Sahelian countries) suggested no different pattern, with lower CHMR at higher levels of IMR clear differences of pattern for the countries of Eastern, and higher CHMR at lower levels of IMR than the other Southern, and Middle Africa, but possibly different pat- areas. terns for the Western Coast and the western Sahel. A final model was estimated relating the log of CHMR to Different symbols are therefore used to identify points in IMR using dummy variables to identify populations of figure 3.5 for the Western Coast, the western Sahel, and the coastal Western Africa and of the western Sahel, plus a other areas. Although there is a considerable amount of dummy variable to identify estimates with a reference peri- scatter, countries of the western Sahel appear in general to od of 1990 or earlier and a dummy variable to identify esti- have higher CHMR relative to IMR than other countries, mates for the period 0 to 4 years before a survey (as opposed but the differences appear to decline as IMR drops. Western to 5 to 9 or 10 to 14 years before a survey). The fitted model Coast countries appear to have broadly similar patterns is as follows: to the countries of the Eastern, Middle, and Southern regions at lower levels of IMR, but lower CHMR at higher ln(CHMR) 2.66 0.0147(IMR) 0.2252(West Coast) levels of IMR. 0.3880(Sahel) 0.1250(1990 or earlier) To explore differences in age patterns more fully, the nat- 0.1167(0­4 years before survey) (3.4) ural log of CHMR was regressed on the IMR for each of three groups of countries: Eastern, Middle, and Southern as one group, Western Coast as another, and western Sahel as As expected from the previous analysis, the Western the third. Predicted values of CHMR for each group for Coast and western Sahel dummy variables are positive and given values of IMR were obtained.3 Figure 3.6 shows the highly significant. The dummy variables for estimates with a summarized age patterns of under-five mortality for the reference date of 1990 or earlier and for the period 0 to 4 three groups. years before the survey are also both positive and significant, although only at a 5 percent level. That the reference date variable is positive implies that, controlling for region and level of IMR, mortality risks between ages one and five fell Figure 3.6 Estimated Relationships between CHMR and IMR, faster in Sub-Saharan Africa in the 1990s than previously, by Region perhaps because of successful measles vaccination pro- (per 1,000 live births) grams. It is puzzling that the coefficient for the dummy vari- 200 able identifying estimates for the five years immediately before the survey is positive. Two errors thought to affect 150 birth histories are a tendency to round age at death up from under one year to "one year," thus tending to inflate CHMR 100 at the expense of IMR, and a tendency to differentially omit CHMR deaths of very young children, thus tending to reduce IMR but without any effect on CHMR. Our prior expectation 50 was that both these errors would be more prevalent the fur- ther before the survey of the event being reported, thus 0 50 100 150 tending to increase CHMR relative to IMR for time periods IMR 5 to 9 or 10 to 14 years before the survey more than in the Countries of the Sahel Western Coast time period 0 to 4 years before the survey. The data suggest Eastern, Middle, and Southern Africa the opposite, but we have no ready explanation for why that Source: Updated version of database compiled by Hill and colleagues (1999). might be so. Trends in Child Mortality, 1960 to 2000 | 21 DISCUSSION increases, along with a "loss of focus" (Jones et al. 2003) on the part of international agencies. Improvements in data sources relevant to the measurement A handful of inexpensive interventions of demonstrated of infant and child mortality in Sub-Saharan Africa have effectiveness could have, with universal coverage, prevented made it possible to use a standard methodology to estimate well over half the 4.4 million deaths of children under age levels and trends in U5MR for most countries over at least five estimated to have occurred in Sub-Saharan Africa in some portion of the last 40 years. This exercise reveals that 2000 (Jones et al. 2003). Efforts to reduce the U5MR Sub-Saharan Africa has performed rather poorly in terms of through the widespread use of these inexpensive interven- child mortality reduction over the period. The average tions may have been responsible for the rapid declines seen annual decline in U5MR of less than 2 percent per year is between 1975 and 1985 in Sub-Saharan Africa and the only about half the average for the developing world overall developing world more broadly. Renewed emphasis on these and less than half the rate needed to meet MDGs. The medi- same interventions, with a small number of recent addi- an U5MR for Sub-Saharan Africa in the year 2000 also tions, such as micronutrient supplementation, can still have exceeded the developing world median by a substantial mar- a major impact today. gin. The countries of Sub-Saharan Africa have performed unevenly over the period in reducing infant and child mor- tality, both by region and by time period. About 1960, coun- APPENDIX 3A: METHODOLOGY tries of Western Africa had high U5MR, on average close to AND DETAIL RESULTS 300 per 1,000 children; countries of Eastern and Southern Africa had somewhat lower risk, on average close to 200 per There are many ways in which a set of estimates can be 1,000; and the countries of Middle Africa had an intermedi- obtained from a series of observations and in which extrap- ate risk, on average about 250 per 1,000. Since 1960, the olations forward or backward to any desired time point can countries of Western and Southern Africa appear to have be made. The simplest procedure is hand smoothing: draw- achieved the most rapid declines in U5MR, and across all ing a freehand curve through a set of observations and regions the period between 1975 and 1985 appears to have extending its general trend onward to some specified time seen the fastest improvements. Rates of decline were lower point for which an estimate or projection is required. Such prior to 1975 and have slowed practically to zero since 1990, a procedure is unlikely to be objective--different analysts and a quarter of all countries actually experienced increases. would almost inevitably draw different lines, particularly for The populations of Sub-Saharan Africa have experienced extrapolations beyond the earliest or latest observations. An changes in many areas other than child survival since 1960. alternative would be to use hand smoothing for the period There have been improvements in literacy and formal covered by observations and then apply a model to extrapo- education, fluctuating but generally rising standards of liv- late; this is essentially what was done in both the United ing, rapid population growth and urbanization, widespread Nations (1988, 1992) reports on under-five mortality in the political instability and civil strife, and since the mid-1980s developing world. With this method, the hand smoothing is the devastating HIV/AIDS epidemic. Further analysis is still subjective, and it is hard to specify the extrapolation required to understand better how all these other changes model satisfactorily. have affected child mortality, why Sub-Saharan Africa has Regression analysis offers a set of possible approaches: performed worse than other developing regions in reducing robust regression, locally weighted least squares, weighted infant and child mortality, why the period 1975 to 1985 was least squares, or ordinary least squares. Such regression tech- the period of most rapid improvement, and why the U5MR niques offer a greater degree of objectivity than hand decline has stalled since 1990. The period of most rapid smoothing but still require the choice of model specifica- improvement, between 1975 and 1985, occurred in a period tion. Each approach has advantages and disadvantages. during which a number of low-cost interventions (oral Ordinary least squares makes few assumptions about the rehydration therapy, vaccines) were developed and distrib- nature of errors, but the results are heavily influenced by uted widely with the support of international agencies and outliers. Robust regression essentially underweights outly- aid programs. The poor performance in the 1990s has coin- ing observations that have an inordinate influence on the cided with rapid expansion of the HIV epidemic, which fitted line, but no prior information about the likely accu- has almost certainly been a factor in the child mortality racy of particular types of observation is used. Locally 22 | Kenneth Hill and Agbessi Amouzou weighted least squares fits a nonparametric trend to a set of Figure 3A.1 Fitted Trends Using Different Numbers of Knot observations, but provides no basis for extrapolation. Parameters Weighted least squares requires the use of predetermined (per 1,000 births) weights, introducing a subjective element into the analysis a. One date and no knot parameters but at the same time allowing the incorporation of prior 400 knowledge about the likely accuracy of different methods of measurement. 5 300 age by Weighted Least Squares with Linear Splines dying 200 of The approach adopted here is that used by Hill and col- 100 leagues (1999). A regression line is fitted to the relationship probability between IMRs or U5MRs and their reference dates using 0 weighted least squares. The basic model assumes that the 1960 1970 1980 1990 2000 2005 rate at which infant or under-five mortality changes is linear date in time, that is, that mortality risk changes at a constant b. One date and two knot parameters annual percentage rate over some defined time period. The 400 dependent variable is the logarithm of either the IMR or the U5MR. The independent variables used are date vari- 5 age 300 ables. The simplest model relates the logarithm of each esti- by mate of the risk of dying to the date of the estimates; this dying 200 model implies a constant rate of change in mortality over of the entire period studied. Figure 3A.1a shows a hypothetical 100 example of such a simple model for under-five mortality. probability More complex models can allow the rate of change of mor- tality to vary over the period; figure 3A.1b is an example of 0 1960 1970 1980 1990 2000 2005 where the relationship between the U5MR and time is date allowed to change two times, reflecting two "knot" points Source: Illustrative country data. (defined every time a sum of weights reaches five, as explained below). Clear changes in trend are indicated in this instance. types (for example, an observation derived from a full birth history for the period zero to four years before the survey) are assigned weights reflecting a priori judgments of their Linear Splines and "Knots" likely validity. The weights for successive observations are The model used in this chapter allows the rate of change of summed, and a knot is defined every time the sum of infant or under-five mortality to vary according to the num- weights reaches a multiple of five. This definition of a knot ber of independent observations available. At the country is based on the idea that five years of vital registration, or level, U5MR is modeled using linear splines, whereby the one set of indirect or direct estimates such as one DHS, are rate of change in U5MR over time is assumed to be constant sufficient to define one trend or slope. The weights assigned for some period, but then is allowed to change at some time to five years of vital registration, or to one set of direct or point or "knot." This model, shown as equation 3A.1, relates indirect estimates, sum to five. Thus, the process of allowing the log of U5MR (by using logs the rate of change is held a knot to be defined when weights sum to five means, in constant) to a date variable and additional variables meas- effect, that each DHS, WFS, or other such survey, and each uring time since a series of knot dates. The number and five-year period of vital registration defines a particular location of these knot dates are determined by the number slope. and location in time of observations available for a particu- The only exception to this procedure is for the last knot lar country. Knots are defined working backward in time defined (the earliest knot in time). For this last knot, the from the most recent observation. Observations of particular remaining weights must sum to at least five to ensure that Trends in Child Mortality, 1960 to 2000 | 23 the first date variable is based on observations whose reported information deteriorates with the length of time weights sum to at least five. The more acceptable observa- elapsed since the events (Som 1973). tions there are for a particular country, the more knots there Each estimate from vital registration or a prospective sur- will be and the more flexible the trend in infant or under- vey is given a "standard" validity weight of 1.0. For vital regi- five mortality over time will be. stration, the weight is justified by the typically large number The underlying model used is of events involved and by the lack of any substantial lag between event and report; for prospective surveys, the ln(nq0)i b0 b1(date)i b2(postk1)i b3(postk2)i weight is justified by the lack of lag and by the accuracy b4(postk3)i . . . ei (3A.1) enforced by the data collection methodology. Estimates The variable date is simply the calendar year; postk1 is derived from birth histories are assigned standard weights date minus the date of the earliest defined knot if positive, that vary with the length of time before the survey to which or zero otherwise, and picks up any change in trend after the the estimate refers, on the grounds that recent information first knot (recall that the knots are defined from the present is more likely to be more accurate than information for peri- backward into the past, but the earliest knot is defined to ods further in the past. Specifically, estimates for the 5 years ensure at least five observations between it and the start of before the survey are given a weight of 2.0; for periods 5 to the series); postk2 is date minus the date of the second 9 years before the survey, 1.8; and for periods 10 to 14 years defined knot if positive, or zero otherwise, and picks up any before the survey, 1.2. change in trend after the second knot; and so on. Thus, the Weights for indirect estimates based on the proportions number of slope-changing time variables varies with the dead of children ever born vary by the age group of the number and weight of the observations over time. The coef- mother; estimates based on reports of young women are ficients on postk1, postk2, and so on, can be interpreted as given low weight--zero for women age 15 to 19 and 0.2 for changes in the rate of change of infant or under-five mor- women age 20 to 24--because of the well-known selection tality with time in that particular period. Thus, the rate of problems that affect such estimates (early childbearing is change in period 1 is b1; in period 2, (b1 b2); in period 3, often highest among the poor, who also suffer the highest (b1 b2 b3); and so on. Referring back to figure 3.1, only U5MRs). Estimates based on reports of women in the age a single date variable with no slope-changing variables groups 25 to 29, 30 to 34, and 35 to 39 are given the highest would be used to fit a trend in U5MR for the Republic of weights, 1.2 each. Then, as age increases the weights decline Congo, since the data weights only sum to five. For Kenya, slowly, on the grounds that information about events longer the date variable and seven slope-changing variables would ago is more prone to error, such that estimates for the age be used, since the data weights sum to 42.5 (figure 3.1b). group 40 to 44 get a weight of 0.8, and those for the age group It should be noted that the error term ei is assumed to be 45 to 49 get a weight of 0.4. normally distributed around the logarithm of the mortality Note that it is these weights that are summed to deter- indicator. As a result, estimates of mortality obtained by mine the location of knots described earlier. The under-five exponentiating an estimated value of the logarithm of the mortality estimates from a particular survey collecting birth mortality indicator will be biased upward by an amount that histories have a combined weight of five, as do the estimates will depend on the goodness of fit of the model. This is a rel- from a particular survey providing indirect measures. In atively benign bias in the sense that the infant or under-five cases in which a particular survey contributes two types of mortality estimates obtained will tend to be on the high side, estimates (for example, a DHS survey that provides both and the poorer the fit of the model the more on the high side direct estimates from the birth history and indirect esti- the estimates will be. mates from children ever born and children surviving), the two sets of weights are reduced to half their standard level, so that the data source is not overly weighted. The Weights observation-specific weights described earlier are essentially Weighted least squares regression is used to fit equation 3A.1 based on the authors' judgment and experience. However, to each country's data. Weighted least squares is used because robust regression techniques can be used to estimate robust a substantial body of evidence suggests different validity weights for particular types of observations; these estimated weights for different types of observations. For example, robust weights can then be compared with the observation- it is generally thought that the quality of retrospectively specific weights.4 24 | Kenneth Hill and Agbessi Amouzou Applying the Methodology obtained. In the process followed in this chapter, this deci- sion is made on the basis of graphical inspection. If the esti- The intention of the methodology is to provide a transpar- mates from one source are clearly higher or lower than the ent and largely objective way of fitting a smoothed trend to bulk of available estimates or if their time trend is clearly a set of observations and of extrapolating the trend to cover different, a constant factor of zero (giving zero weight to the the period from 1960 to the present. However, subjective entire data set) might be used. If the estimates fluctuate a lot judgments are involved at every step. or display some other undesirable characteristic a constant In step 1 of the smoothing and extrapolation process, the factor between zero and one might be used, thus reducing regression model in equation 3A.1 is fitted to observations the standard weight for the entire data set. It is this decision of U5MR using appropriate date variables and standard that is most likely to introduce substantial differences in validity weights as defined above. Results from step 1 results between analysts. In the analysis reported here, it is depend on the standard validity weights applied to different assumed that errors are more likely to result in underesti- types of data. The standard validity weights are objective in mates of under-five mortality than in overestimates. Thus, the sense that they are invariant across data sets or across when two data sets indicate very different levels, other things countries, but they are subjective in the sense that different being equal, the set indicating higher mortality is assumed analysts would choose different values. more likely to be right. In step 2, results from step 1 are critically examined and data sets that are clearly aberrant are identified, such as indi- rect estimates that are clearly inconsistent with the majority APPENDIX 3B: ESTIMATES OF U5MRs AND IMRs of other sources. The weights for the entire aberrant data set AND DATA SOURCES are reduced by a constant factor that is generally zero (giving no weight to estimates from that data set). Step 1 is then Tables 3B.1 and 3B.2 present the estimates of U5MRs and repeated using these revised weights. IMRs, respectively. Countries are categorized into four This second step involves subjective choices as well. Most groups depending on the number of surveys available and important is the decision as to whether to underweight the consistency of the under-five mortality estimates entire data sets and, if so, by how much. The decision to obtained from those surveys. Data sources for each country underweight some data sets relative to others can have a large are given in table 3B.3. Figure 3B.1 shows an example of a effect on the sequence of under-five mortality estimates country in each category. Trends in Child Mortality, 1960 to 2000 | 25 Table 3B.1 Estimates of Under-Five Mortality, by Country and Year (per 1,000 births) No. of censuses Under-five mortality rate and Region and country surveys Category 1960 1965 1970 1975 1980 1985 1990 1995 2000 Middle Angola 2 C -- -- -- -- 262 261 260 255 249 Cameroon 3 C 258 235 214 196 172 150 141 156 173 Central African Republic 3 C 340 286 240 202 189 177 164 152 141 Chad 3 C -- -- -- 233 214 196 195 209 226 Congo, Democratic Republic of 1 C 276 257 239 222 207 202 202 202 202 Congo, Republic of 2 C 197 166 139 -- -- -- -- -- -- Gabon 1 C -- -- -- -- 104 99 95 91 87 São Tomé and Principe 1 D -- -- -- -- 112 115 120 124 128 Eastern Burundi 5 A 248 237 227 214 190 186 183 192 201 Comoros 2 D -- -- -- 174 154 136 121 107 94 Eritrea 1 C 206 177 152 131 112 Ethiopia 4 B 259 248 238 228 218 209 201 193 185 Kenya 9 A 204 179 158 139 115 104 97 114 134 Madagascar 3 C 138 141 149 159 170 181 174 147 125 Malawi 7 A 361 351 330 296 265 247 233 210 189 Mozambique 3 D 309 292 276 261 247 233 221 208 197 Rwanda 5 A 206 190 209 226 216 188 168 192 220 Somalia 1 C -- -- -- -- 195 203 211 219 227 Tanzania 7 B 240 228 218 203 175 157 157 151 142 Uganda 5 A 223 194 187 186 185 180 160 153 150 Zambia 5 A 213 195 181 162 150 166 184 204 227 Zimbabwe 8 D 158 149 139 127 109 84 74 94 123 Southern Botswana 6 B 173 159 147 118 95 71 63 65 68 Lesotho 4 B 203 200 190 173 158 144 131 120 109 Namibia 3 C 244 200 164 134 110 90 74 59 47 South Africa 1 C -- -- -- -- 86 74 64 55 48 Swaziland 4 A 224 217 196 172 143 119 108 108 108 Western Benin 3 C 298 274 252 232 214 199 184 170 156 Burkina Faso 4 A 315 296 278 261 242 223 211 223 236 Côte d'Ivoire 5 A 349 286 235 192 172 156 158 179 203 Gambia, The 4 A 367 340 315 269 227 192 164 143 125 Ghana 5 B 214 201 188 171 157 141 122 109 97 Guinea 2 C -- -- -- 335 297 263 232 188 153 Guinea-Bissau 1 D -- -- -- -- 286 267 250 233 218 Liberia 4 B 284 276 269 252 235 219 204 191 178 Mali 4 A 518 451 392 341 297 265 252 240 229 Mauritania 3 D 223 213 205 196 188 181 182 187 193 Niger 3 D -- -- -- 287 291 296 293 273 249 Nigeria 3 D 279 255 234 214 196 179 165 153 143 Senegal 5 A 297 288 279 265 216 175 150 148 148 Sierra Leone 4 A 390 393 370 346 321 297 286 300 314 Togo 3 C 267 240 215 194 175 158 152 148 143 Source: Authors' calculations. Notes: -- no observed data available for that year or for years around it. The "Category" column indicates a subjective judgment of the quality of the estimates available for a given country. The cate- gories are as follows: A four or more surveys with broadly consistent results; B four or more surveys with broadly inconsistent results; C three or fewer surveys with broadly consistent results; D three or fewer surveys with broadly inconsistent results. 26 | Kenneth Hill and Agbessi Amouzou Table 3B.2 Estimates of Infant Mortality, by Country and Year (per 1,000 births) Infant mortality rate Region and country 1960 1965 1970 1975 1980 1985 1990 1995 2000 Middle Angola -- -- -- -- 156 155 154 152 150 Cameroon 153 139 127 117 104 93 88 96 105 Central African Republic 194 166 145 127 121 115 109 103 98 Chad -- -- -- 138 127 117 117 124 133 Congo, Democratic Republic of 162 153 145 137 130 128 128 127 127 Congo, Republic of 133 114 97 -- -- -- -- -- -- Gabon -- -- -- -- 75 72 69 67 64 São Tomé and Principe -- -- -- -- 80 82 85 87 90 Eastern Burundi 147 140 134 127 114 112 110 115 120 Comoros -- -- -- 118 106 95 85 77 69 Eritrea 138 120 105 92 80 Ethiopia 173 166 159 152 146 140 135 130 126 Kenya 122 108 97 87 73 67 63 72 83 Madagascar 86 87 92 97 103 109 105 91 78 Malawi 205 200 189 171 157 149 142 131 121 Mozambique 178 169 162 155 149 142 136 131 125 Rwanda 122 114 124 134 128 113 102 115 130 Somalia -- -- -- -- 117 121 125 130 135 Tanzania 142 135 129 121 106 96 96 93 88 Uganda 132 116 112 112 111 109 98 94 92 Zambia 126 117 109 99 92 101 111 121 134 Zimbabwe 97 92 86 80 70 56 50 61 78 Southern Botswana 118 110 102 84 69 54 48 50 52 Lesotho 136 135 128 118 109 100 92 85 78 Namibia 147 126 109 94 82 70 59 49 40 South Africa -- -- -- -- 63 56 49 43 38 Swaziland 150 145 132 117 100 84 77 77 77 Western Benin 177 162 149 138 127 118 111 103 96 Burkina Faso 181 171 163 155 146 137 132 137 143 Côte d'Ivoire 236 192 157 130 118 107 109 122 136 Gambia, The 208 194 181 159 139 123 109 99 90 Ghana 127 119 113 104 96 87 77 69 63 Guinea -- -- -- 191 172 156 142 121 104 Guinea-Bissau -- -- -- -- 191 179 167 156 146 Liberia 190 185 180 168 157 147 137 129 121 Mali 293 254 222 194 172 157 151 146 140 Mauritania 149 143 137 132 127 123 123 127 130 Niger -- -- -- 170 173 176 174 162 147 Nigeria 166 151 138 127 117 108 100 94 88 Senegal 172 167 163 157 134 114 102 101 101 Sierra Leone 220 222 210 197 184 172 167 173 180 Togo 158 142 128 116 106 97 94 91 89 Source: Authors' calculations. Notes: -- no observed data available for that year or for years around it. Trends in Child Mortality, 1960 to 2000 | 27 Table 3B.3 Countries and Data Sources Region and country Data sources Middle Angola MICS, 2000 Cameroon WFS, 1978; DHS, 1991, 1998 Central African Republic Census, 1975, 1988; DHS, 1995 Chad Census, 1993; DHS, 1997; MICS, 2000 Congo, Democratic Republic of Census, 1984; MICS, 1996 Congo, Republic of Census, 1974 Gabon DHS, 2001 São Tomé and Principe MICS, 2000 Eastern Burundi DS, 1970; PCS, 1979; DHS, 1987; Census, 1990; MICS, 2000 Comoros DHS, 1996; MICS, 2000 Eritrea DHS, 1995 Ethiopia DS, 1981; Census, 1984; FFS, 1990; DHS, 2000 Kenya Census, 1960; DS, 1977; WFS, 1977; Census, 1979; DS, 1983, 1988, 1993; Census,1989; DHS, 1998 Madagascar Vital Reg., 1960­61, 1963­72; DS, 1966; DHS, 1992, 1997 Malawi PCS, 1970; Census, 1977; MDS, 1982; FFS, 1984; Census, 1987; DHS, 2000, 1992 Mozambique Census, 1970, 1980; DHS, 1997 Rwanda DS, 1970, 1978; NFS, 1983; DHS, 1992, 2000 Somalia MICS, 2000 Tanzania Census, 1967; DS, 1973; Census, 1978, 1988; DHS, 1991, 1994, 1996, 1999 Uganda Census, 1969; DHS, 1988; IHS, 1992; DHS, 1995, 2000 Zambia Census, 1969; SCP, 1974; Census, 1980; DHS, 1992, 1996 Zimbabwe Census, 1969, 1982; RHS, 1984; ICD, 1987; DHS, 1988; Census, 1992; DHS, 1994, 1999 Southern Botswana Vital Reg., 1982­87; Census, 1971, 1981; DHS, 1984, 1988; Census, 1991; MICS, 2000 Lesotho CES, 1968; DS, 1971; Census, 1976; WFS, 1977 Namibia DHS, 1992, 2000 South Africa DHS, 1998 Swaziland Census, 1966, 1976, 1986; MICS, 2000 Western Benin WFS, 1982; DHS, 1996, 2001 Burkina Faso PES, 1976; Census, 1985; DHS, 1992, 1999 Côte d'Ivoire MRS, 1978; WFS, 1980; DHS, 1994, 1998 Gambia, The Census, 1973, 1983, 1993; MICS, 2000 Ghana Census, 1971; FS, 1979; DHS, 1988, 1993, 1998 Guinea DHS, 1992, 1999 Guinea-Bissau MICS, 2000 Liberia PGS, 1969, 1970; Census, 1974; DHS, 1986 Mali Census, 1976; DHS, 1987, 1995, 2001 Mauritania WFS, 1981; MICS, 1996; DHS, 2000 Niger DHS, 1992, 1998; MICS, 2000 Nigeria WFS, 1981; DHS, 1990, 1999 Senegal WFS, 1978; DHS, 1986, 1992, 1997, 1999 Sierra Leone Pilot Census, 1973; Census, 1974, 1985; MICS, 2000 Togo DS, 1971; DHS, 1988, 1998 Source: Authors' compilation. Note: CES Consumption and Expenditure Survey; DS Demographic Survey; FS Fertility Survey; FFS Family Formation Survey; ICD Inter-Censal Demographic Survey; IHS Integrated Household Survey; MRS Multiple Round Survey; NFS National Fertility Survey; PCS Population Change Survey; PES Population Enumeration Survey; PGS Population Growth Survey; RHS Retrospective Health Survey; SCP Sample Census of Population; Vital Reg. vital registration. 28 | Kenneth Hill and Agbessi Amouzou Figure 3B.1 Examples of "Quality" Categories for Selected Countries (per 1,000 births) Category A: Kenya Category B: Ghana 300 300 5 250 5 250 age age by 200 by 200 dying 150 dying 150 of of 100 100 probability 50 probability 50 0 0 1960 1970 1980 1990 2000 2005 1960 1970 1980 1990 2000 2005 year year Census WFS/DHS direct WFS/DHS indirect Census DHS/WFS direct DHS/WFS indirect Fitted line NDS indirect Fitted line Category C: Guinea Category D: Mauritania 500 300 5 5 250 400 age age by by 200 300 dying dying 150 of of 200 100 probability 100 probability 50 0 0 1960 1970 1980 1990 2000 2005 1960 1970 1980 1990 2000 2005 year year DHS direct DHS indirect Fitted line DHS/WFS indirect MICS indirect Fitted line Source: Updated version of database compiled by Hill and colleagues (1999). NOTES Africa: ln(CHMR) 3.90 0.0049(IMR); West Sahelian Africa: ln(CHMR) 3.46 0.0122(IMR). 1. Theoretically, in a population with perfect registration of vital events 4. See Hill et al. (1999) for an example of use of robust regression to and no migration, the mortality of children after infancy could also be estimate the weights. measured purely from civil registration data. However, in practice, regis- tration is not perfect, and migration is not negligible, so mortality in child- hood is calculated from deaths from civil registration data and from expo- sure to risk according to a census or population estimate. REFERENCES 2. A box plot summarizes the distribution of a variable. The box is a rectangle, the top and bottom of which mark the 75th and 25th per- Brass, W. 1964. "Uses of Census or Survey Data for the Estimation of Vital centiles, respectively, with the median observation (in this case, the medi- Rates." E/CN.14/CAS.4/V57. Paper presented at the African Seminar an country) as a cross-bar within the box. The "whiskers" for each box are on Vital Statistics, Addis Ababa, December 14­19. the lines protruding above and below, and indicate the range of the data above and below the upper and lower quartiles. Coale, A. J., and P. Demeny. 1983. Regional Model Life Tables and Stable 3. The robust regression coefficients for the regions were as follows (all Populations. 2nd ed. New York: Academic Press. coefficients were significant at the 0.001 level or better). Eastern, Middle, Hill, K. 1991. "Approaches in the Measurement of Childhood Mortality: A and Southern Africa: ln(CHMR) 2.50 0.0179(IMR); coastal West Comparative Review." Population Index 57 (3). Trends in Child Mortality, 1960 to 2000 | 29 ------. 1995. "Age Patterns of Under-Five Mortality in the Developing Sullivan, J. M. 1972."Models for the Estimation of the Probability of Dying World." Population Bulletin of the United Nations 39. between Birth and Exact Ages of Early Childhood." Population Studies Hill, K., R. Pande, M. Mahy, and G. Jones. 1999. Trends in Child Mortality 26 (1): 79­97. in the Developing World: 1960­1996. New York: UNICEF. Trussell, T. J. 1975. "A Re-estimation of the Multiplying Factors for the Hill, K., and A. Yazbeck. 1994. Trends in Under-Five Mortality, 1960­90: Brass Technique for Determining Childhood Survivorship Rates." Estimates for 84 Developing Countries. Washington, DC: World Population Studies 29 (1): 97­108. Bank. United Nations. 1983. Manual X: Indirect Techniques for Demographic Estimation. Population Studies 81. New York: United Nations. Jones, G., R. Steketee, R. Black, Z. Bhutta, S. Morris, and the Bellagio Child Survival Study Group. 2003."How Many Child Deaths Can We Prevent ------. 1988. Mortality of Children under Age 5. World Estimates and This Year?" Lancet 362: 65­71. Projections, 1950­2025. Population Studies, No. 105. Department of International Economic and Social Affairs. New York: United Som, R. K. 1973. Recall Lapse in Demographic Surveys. Bombay: Asia Nations. Publishing House. ------. 1992. Under-Five Mortality since the 1960s: A Database for Somoza, J. L. 1980. Illustrative Analysis: Infant and Child Mortality in Developing Countries. New York: United Nations. Colombia. World Fertility Survey Scientific Reports 10. London: World Fertility Survey. 30 | Kenneth Hill and Agbessi Amouzou Chapter 4 Levels and Trends of Adult Mortality Debbie Bradshaw and Ian M. Timaeus Adult mortality remains a neglected public health issue in related to underdevelopment, such as malaria, malnutrition, Sub-Saharan Africa. A lack of empirical data about the lev- and tuberculosis. In addition, they highlight the increasing els of mortality experienced by adults in this region has road traffic injury burden and the less predictable toll of fueled this neglect, combined with the focus on maternal mortality due to war and violence. and child health, which has the highest incidence of disease Health sector reforms have swept the continent during and subsequent mortality. This picture is changing. The the last 20 years and given primary health care priority in high mortality of adults in the African region is now being attempts to provide accessible health care to rural commu- recognized more widely, and a response has begun to nities and contain costs. However, the strategies of primary emerge, particularly with regard to the impact of the AIDS health care have focused largely on maternal and child epidemic and high mortality due to malaria. health and the provision of acute care. These strategies have The Global Burden of Disease studies of 1990 and 2000 yet to overcome the challenges posed by the health needs of estimate that Sub-Saharan Africa has the highest burden of adults, such as the management of chronic conditions and, disease in the world (Murray and Lopez 1996; WHO 2000). more recently, the provision of treatment for AIDS. The These and related studies have revealed high levels of adult imperative of these challenges has been obscured by the lack mortality resulting from the multiple burdens of disease of good data on the levels of adult mortality and its trends experienced by the populations in the region (Murray, Yang, and causes. Although childhood and maternal mortality and Qiao 1992). The studies show that fast-growing epi- once faced the same problem, survey-based techniques, demics of human immunodeficiency virus and acquired including indirect methods of estimation and the wide- immune deficiency syndrome (HIV/AIDS) and certain spread collection of birth histories, have improved the situ- noncommunicable diseases coexist with the conditions ation in this regard. 31 SOURCES OF DATA and Hill (1997) suggest that sibling survival may progres- sively underestimate adult mortality as the time since the Vital registration systems are the ideal source of data with event lengthens. These authors conclude that sibling histories which to monitor mortality levels and trends. However, the can at best provide an estimate of adult mortality in the few statistics systems of most Sub-Saharan Africa countries con- years before the data were collected. Comparison of succes- tinue to be underdeveloped; few countries are able to pro- sive sets of sibling history data in the three African countries vide adequate vital statistics. In 2001 the World Health that have collected them in two DHS surveys also supports Organization (WHO) contacted all member states in Sub- this conclusion (Timaeus and Jasseh 2004). Apart from the Saharan Africa to try to obtain death statistics for a global DHS surveys, several other international programs of sur- assessment of the burden of disease (Kowal, Rao, and veys have been conducted in Sub-Saharan Africa during the Mathers 2003). Just 9 out of 46 member states provided such past decade. However, although the Multiple Indicator data. Furthermore, coverage of registration was less than Cluster Surveys (MICS) conducted by UNICEF are an 60 percent in most of these 9 countries, with only Mauritius important source of child mortality estimates, they have not and the Seychelles providing vital registration data that were included questions on adult mortality. Moreover, data from more than 90 percent complete. These two island states the World Health Survey have yet to become available. make up a tiny fraction of the population of Sub-Saharan Data on adult mortality can be collected in national cen- Africa. Adequate vital statistics are far more widespread in suses either by including questions about recent deaths in other regions, particularly Europe and the Americas, and households or by asking about the survival of the parents of 60 percent of the countries for which the WHO could not household members. Only some African countries have a obtain registration data are in Sub-Saharan Africa. Data tradition of asking one or both sets of questions in their cen- series of any time length are particularly rare. Due to a lack suses, and few additional countries responded to the grow- of appropriate direct data, the Adult Mortality in ing concern about adult mortality from AIDS and other Developing Countries project, which studied trends in adult causes by adding them to their schedule for the 2000 census. mortality in 27 developing countries (Hill 2003), included Even fewer countries have yet published these data. only 2 countries from mainland Sub-Saharan Africa (Benin Although reporting of recent deaths can often be incom- and Zimbabwe). Among the mainland countries, South plete and subject to reference period and age reporting Africa is an exemplary country because of the government's errors, a range of methods exists for the evaluation of such determined efforts in the last decade to improve its statistics. information and, in favorable circumstances, the correction The lack of reliable mortality data had been dubbed the of inadequate data (Bennett and Horiuchi 1984; Brass 1975; "black hole of vital statistics" (Botha and Bradshaw 1985), Hill 1987). Data on the survival of parents can be used to but national efforts during the 1990s resulted in over 90 per- estimate conventional life table indexes by means of the cent of the adult deaths being registered (Dorrington et al. orphanhood method (Brass and Hill 1973; Timaeus 1992). 2001). Yet, even in South Africa, the production of timely The method can be used to estimate the historical trend in cause-of-death statistics remains a challenge. mortality and has been refined in various ways in order to The pervasive lack of vital registration data makes it nec- produce more up-to-date estimates (for example, Timaeus essary to derive estimates of mortality using indirect demo- 1991). Moreover, adjustments can be made to orphanhood graphic techniques based on survey and census data. The data to correct for the biases introduced by the excess mor- most readily accessible data that can be used to estimate adult tality of orphans who have been vertically infected with HIV mortality come from the Demographic and Health Surveys (Timaeus and Nunn 1997). (DHS). At least one such survey has been conducted in most Two consecutive censuses can also be used to estimate African countries, and many of the African surveys include adult mortality from intercensal cohort survival or age- sibling histories that were developed to obtain estimates of specific growth (Preston and Bennett 1983). The ratio maternal mortality (Rutenberg and Sullivan 1991). resulting from matching age groups in the second popula- Unfortunately, almost all DHS samples are too small to tion to the first population provides a summary measure of enable the study of age-specific mortality in adulthood or adult mortality. This requires that the population be closed mortality trends in any detail. Moreover, concerns exist to migration and that the census coverage be the same in about both the reliability and validity of reports on the sur- both censuses. Few, if any, Sub-Saharan Africa countries vival of respondents' siblings. Although Bicego (1997) found come close to meeting either of these conditions. Thus, the good correspondence with other data, Stanton, Abderrahim, method has not been used much or with great success. 32 | Debbie Bradshaw and Ian M. Timaeus Demographic surveillance of local populations has been The WHO reviewed the various estimates and observed used for many years in a handful of African populations to substantial variations, depending on the different proce- study mortality and provide a framework for epidemiologi- dures and judgments made (Lopez et al. 2002). It produced cal research and the trial of health interventions (for exam- its own set of life table estimates for 191 countries in 2000 ple, Pison and Langaney 1985). In response to the spread of based on its own mortality database and a new family of HIV in Africa, many more such longitudinal surveillance model life tables, which in turn were based on an extension systems have been established recently in which vital events of the Brass logit system (Murray et al. 2003). Although the are monitored at regular intervals (Gregson et al. 1997; basis of these life tables was measures of mortality from Hosegood, Vanneste, and Timaeus 2004; Nunn et al. 1997; direct registration wherever possible, the only mainland Sewankambo et al. 2000; Todd et al. 1997; Tollman et al. Sub-Saharan Africa countries for which such data were 1999; Urassa et al. 2001). Almost all these sites, however, adopted were Zimbabwe and South Africa and in addition, cover only small and select populations. Therefore, they are for Tanzania, where the AMMP surveillance data for part of of little use for the estimation of adult mortality in national the country were used. populations. The partial exception is the surveillance system For the rest of Africa, the WHO, like the UN Population organized by the Adult Morbidity and Mortality Project Division, was forced to estimate adult mortality and life (AMMP) in Tanzania, which covers three districts that expectancy by extrapolation from estimates of child mortal- include both rural and urban areas, although it is not statis- ity. It did this by modeling the relationship between adult tically representative of the population of the country as a and child mortality in its database of more than 1,800 life whole (Kitange et al. 1996). tables, using its modified logit life table system. Given the Because of the shortage of reliable empirical estimates, lack of such data, however, this database includes few life World Population Prospects (United Nations 2003) uses esti- tables from Sub-Saharan Africa and few with levels of child mates of child mortality and assumptions about the age pat- mortality as high as those that are typical of the African tern of mortality taken from a family of model life tables to region. Thus, although the WHO's estimates are based on determine life expectancy in all of mainland Sub-Saharan models developed using contemporary developing country Africa. The system of model life tables adopted by the data, there is no guarantee that they will be more accurate United Nations (UN) Population Division implicitly defines for Africa than those made by the UN Population Division the level of adult mortality. In most of these countries the using model life tables based on historical data on high- Population Division simply has assumed that non-AIDS mortality Western populations. Moreover, like the mortality at all ages conforms to a Princeton North model Population Division, the WHO was unable to locate any up- life table. These models are derived from historical data to-date mortality data on some eight African countries. To largely on northern Europe (Coale, Demeny, and Vaughan produce the final life tables, mortality levels were estimated 1983) and are fitted to an estimate of under-five mortality without HIV/AIDS, and then estimates of the additional (United Nations 2002). In a few countries, Princeton West adult AIDS deaths were added, based on the impact models (Coale, Demeny, and Vaughan 1983) or UN Far observed in the direct estimates from the demographic sur- Eastern models (United Nations 1982) have been used in the veillance sites in Tanzania and the national vital registration same way. Moreover, Sub-Saharan Africa is the only area in systems of Zimbabwe and South Africa (Salomon and which data are completely lacking for a few countries and Murray 2001). mortality in them has to be guessed from estimates for neighboring countries. The UN Population Division bases its estimates of AIDS deaths in high-prevalence African MORTALITY LEVELS countries on those produced by the Joint United Nations Programme on HIV/AIDS (UNAIDS) using epidemiologi- Table 4.1 shows estimates of adult mortality based on the cal models fitted to HIV seroprevalence data collected in WHO life tables for Africa in 2000 (Lopez et al. 2002). The antenatal clinics (UNAIDS Reference Group 2002). index presented is the probability of dying between exact Although these models are based on careful review of the ages 15 and 60 (45q15). It refers to a particular year but can data coming out of a range of research studies, at present no be interpreted as the probability that someone who had attempt is made by any of the UN agencies to use empirical survived childhood would die before old age if he or she data on mortality in national populations to estimate AIDS went through life subject to the age-specific death rates of deaths in particular countries. the year in question. This measure has been adopted by Levels and Trends of Adult Mortality | 33 Table 4.1 Probabilities of Dying between Ages 15 and 60 (45q15) in Sub-Saharan Africa WHO region and country Male Female WHO region and country Male Female Southern Africa Western Africa Angola 0.492 0.386 Algeria 0.155 0.119 Botswana 0.703 0.669 Benin 0.384 0.328 Lesotho 0.667 0.630 Burkina Faso 0.559 0.507 Malawi 0.701 0.653 Cape Verde 0.210 0.121 Mozambique 0.674 0.612 Côte d'Ivoire 0.553 0.494 Namibia 0.695 0.661 Gambia, The 0.373 0.320 South Africa 0.567 0.502 Ghana 0.379 0.326 Swaziland 0.627 0.587 Guinea 0.432 0.366 Zambia 0.725 0.687 Guinea-Bissau 0.495 0.427 Zimbabwe 0.650 0.612 Liberia 0.448 0.385 Eastern Africa Mali 0.518 0.446 Burundi 0.648 0.603 Mauritania 0.357 0.302 Eritrea 0.493 0.441 Niger 0.473 0.408 Ethiopia 0.594 0.535 Nigeria 0.443 0.393 Kenya 0.578 0.529 Senegal 0.355 0.303 Rwanda 0.667 0.599 Sierra Leone 0.587 0.531 Tanzania 0.569 0.520 Togo 0.460 0.406 Uganda 0.617 0.567 Indian Ocean Central Africa Comoros 0.381 0.325 Cameroon 0.488 0.559 Madagascar 0.385 0.322 Central African Republic 0.620 0.573 Mauritius 0.228 0.109 Chad 0.449 0.361 Seychelles 0.268 0.122 Congo, Dem. Rep. of 0.571 0.493 Congo, Rep. of 0.475 0.406 Equatorial Guinea 0.339 0.280 Gabon 0.380 0.330 São Tomé and Principe 0.269 0.226 Source: Authors' calculations; data from Lopez et al. 2002. many international agencies, including the WHO, as a In its estimates of mortality, the WHO found that of the good indicator of the overall level of adult mortality. Unlike 40 countries with the highest mortality, 37 were from the life expectancy at age 15, it is not influenced by death rates Sub-Saharan region. The level of adult mortality is highly at age 60 and over, which are very difficult to measure in variable across African countries. Southern and Eastern countries with defective vital statistics systems. One limita- Africa have particularly high adult mortality, whereas mor- tion of the measure is that it is heavily influenced by death tality in Western Africa is lower and the Indian Ocean rates in the upper part of the 45-year age range. This Islands, which accommodate relatively small populations, reduces its sensitivity to trends in death rates in early adult- have the lowest rates. According to these WHO estimates, hood. In addition, without supplementary data or assump- the probability of surviving from exact age 15 to exact age 60 tions, neither sibling history data, which refer mainly to in 2000 was less than 50 percent in nearly half of the young adults, nor orphanhood data, which refer largely to countries in Sub-Saharan Africa. Most of these are countries middle-aged parents, can provide robust measures of adult affected severely by the HIV/AIDS epidemic, but the list mortality across the whole of this wide age range. also includes war-torn countries, such as Sierra Leone. The 34 | Debbie Bradshaw and Ian M. Timaeus Figure 4.1 Probabilities of Dying between Exact Ages 15 and TRENDS IN ADULT MORTALITY 60 in Sub-Saharan Africa, by Sex There is growing evidence of rising trends in adult mortality 0.8 in the countries in Sub-Saharan Africa. The eastern and 0.7 southern African regions have been particularly hard hit 0.6 estimates by the AIDS epidemic, and the available data show large 0.5 Namibia Botswana increases in adult mortality rates. For example, data from Lesotho Division 0.4 Namibia Malawi show that adult mortality was declining until the Botswana 0.3 South Africa mid-1980s but that this trend was reversed in the 1990s 0.2 Population (table 4.2). These results are based on the intercensal sur- UN 0.1 vival method but are broadly consistent with estimates 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 based on recent household deaths and DHS sibling history WHO estimates data (Blacker 2004). They suggest that mortality for men and women were at similar levels, rather than women hav- men women ing lower mortality, but long-standing evidence exists to Source: Authors' calculations from United Nations 2003 and Lopez et al. 2002. suggest that Malawi may be exceptional in this regard Note: WHO estimates are for 2000; UN Population Division estimates are for 1995­2000. (Timaeus 1993, 1998). Analyses of data from censuses in Kenya also suggest that estimates suggest that adult men have consistently higher a reversal in the trend in adult mortality occurred in that mortality than adult women in Sub-Saharan Africa, with the country during the 1990s (table 4.3). The orphanhood female advantage in survivorship being smallest in the very estimates of person-years lived indicate that slight high mortality countries and largest in the low-mortality improvements were made between the 1970s and 1980s but island states. that an even larger decline occurred in the next decade. In The estimates of adult mortality produced by the WHO this country, the mortality of women is lower than that of for 2000 and those made by the United Nations Population men. Division for the period 1995­2000 are compared in fig- ure 4.1. It is clear that the WHO estimates tend to be signif- icantly higher than those issued by the Population Division Table 4.2 Probabilities of Dying between Ages 15 and 60 for both men and women. The discrepancies are especially (45q15), by Sex, according to Intercensal Survival, Malawi large in several countries in the Southern African region, but are also significant for men and of borderline significance Period Males Females for women (p 0.059) in the other countries of the region. 1966­77 0.391 0.344 As much of Africa is experiencing rising adult mortality 1977­87 0.243 0.290 from HIV/AIDS, one would expect the WHO's 2000 esti- 1987­98 0.487 0.429 mates to be slightly higher than those made by the Source: Blacker 2004. Population Division for a period on average 2.5 years earli- er. However, HIV/AIDS can explain only a small part of the discrepancies between the two sets of estimates. Instead, Table 4.3 Orphanhood Estimates of Person-Years Lived and they are rooted in part in differences in what is assumed Lost at Ages 35­75 (Men) and 25­75 (Women), Kenya about the level of child mortality by the two organizations, in part in the use of different model life tables to extrapolate Hypothetical Intercensal Cohorts from child to adult mortality, and in part in different 1969­79 1979­89 1989­99 assumptions about the scale of AIDS mortality in particular Males 35­75 countries. One cannot readily quantify the relative Years lived 31.0 32.5 29.6 importance of these contributions to the scatter revealed Years lost 9.0 7.5 10.4 by figure 4.1. It seems likely, however, that the large discrep- Females 25­75 ancies between the estimates for some southern African Years lived 43.4 44.8 42.8 countries result at least in part from different assumptions Years lost 6.6 5.2 7.2 about the severity of AIDS mortality in them. Source: Blacker, Kizito, and Obonyo 2003. Levels and Trends of Adult Mortality | 35 Analysis of a range of data sources from Zimbabwe, Figure 4.3 Conditional Probabilities of Dying between Ages including vital registration data, has also shown a marked 15 and 60 in South Africa, by Sex, from Different Data rise in the adult mortality rates during the 1990s (Feeney Sources, 1980­2000 2001). This is in keeping with the rising prevalence of HIV 0.5 observed in antenatal clinics during the late 1980s and early 60 1990s. Trends in the probability of dying between ages 30 and 0.4 15 and 65, 35 30 q , are shown in figure 4.2. The estimates from vital registration, household deaths, and survival of parents 0.3 between are reasonably consistent, although the last of these series may underestimate somewhat for the earlier period and dying 0.2 of place the subsequent rise in mortality slightly too early. Mortality probably fell during the 1970s and early 1980s, but 0.1 probability age-specific death rates and probabilities of death subse- 0 quently rose by 200 to 300 percent between the late 1980s 1980 1985 1990 1995 2000 and late 1990s for both adult men and adult women. year Analysis of a range of data sources from South Africa that women--registration LSDS--mothers OHS 1995--mothers also include vital registration data shows that adult mortal- men--registration LSDS--fathers OHS 1995--fathers census--mothers DHS--mothers OHS 1998--mothers ity rates were fairly constant during the 1980s, followed by census--fathers DHS--fathers OHS 1998--fathers DHS--sisters (raw) DHS--sisters (modeled) OHS women--direct an increase in the late 1990s (Timaeus et al. 2001). Trends in DHS--brothers (raw) DHS--brothers (modeled) OHS men--direct 45 15 q , the probability of surviving from 15 to 60, are shown in figure 4.3 and reveal the initial signs of the impact of the Source: Timaeus et al. 2001. Note: DHS Demographic and Health Survey, 1998; LSDS Living Standards and Development major HIV epidemic that emerged during the 1990s. More Study, 1993; and OHS October Household Study. recent data collated from the national population register show that there has been a continued rise in the number of adult deaths registered in South Africa (figure 4.4) and the account for some of this increase. Taking into account these increase has followed the distinct age pattern associated with improvements as well as population growth, the increase in AIDS. However, improved registration of deaths and deaths of persons at ages 15 years or more has been approx- improved registration on the population register could imately 40 percent. Age-specific rates can be calculated for South Africa from these data by adjusting for underregistra- tion of deaths (Dorrington, Moultrie, and Timaeus 2004). Figure 4.2 Conditional Probabilities of Dying between Ages This can only be done for the years up to 2001, the date of 30 and 65 in Zimbabwe, by Sex, from Different Data Sources, the most recent census, which provides an estimate of the 1969­90 population against which to assess the underregistration. It 0.8 confirms that a large rise in the mortality of young adults 65 0.7 has occurred. The relative increase in mortality rates has and 30 0.6 been greater for women than for men, although starting 0.5 from a lower base. Also, the increase for women occurs at a between 0.4 younger age than that for men. dying Comparison of the empirically based estimates presented 0.3 of in figure 4.3 with those in table 4.1 suggest that the model- 0.2 ing procedure used by the WHO to estimate adult mortality 0.1 probability in South Africa in 2000 has produced severe overestimates, 0 1960 1970 1980 1990 2000 particularly for women. This highlights the uncertainty year attached to estimates of adult mortality made from data women--death registration men--death registration on children and to projections using epidemiological mod- women--household death men--household death women--mothers' survival, 1982 els fitted to HIV data. It may be the case that the WHO men--fathers' survival, 1982 women--mothers' survival, 1992 men--fathers' survival, 1992 also overestimated women's mortality in the other south- women--mothers' survival, 1997 men--fathers' survival, 1997 ern African countries for which it published much higher Source: Feeney 2001. estimates than the UN (figure 4.1). However, the estimates 36 | Debbie Bradshaw and Ian M. Timaeus Figure 4.4 Age Distribution of Reported Adult Deaths on the which is concentrated among young adults, as in the national National Population Register of South Africa, by Sex, statistics, can be accounted for entirely by AIDS deaths. 1998­2003 Church records from northern Namibia have been used a. Registered adult male deaths to estimate the mortality experience of the parishioners 28,000 (Notkola, Timaeus, and Siiskonen 2004). Adult mortality 24,000 did not change much during the 1980s but increased rapid- ly from about 1994. In 2000 the death rates of women at ages 20,000 20­64 were 3.5 times higher than in 1993; for men they were deaths 16,000 of 2.5 times higher. 12,000 An analysis of the sibling histories collected prior to 1997 number 8,000 in 11 DHS surveys in Sub-Saharan Africa showed that, 4,000 although adult mortality was falling or stagnant in Western Africa and in Namibia in the 1980s, it had begun to rise 0 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95+ sharply in Eastern Africa (Timaeus 1998, 1999). Moreover, age four of the six Eastern African countries considered were b. Registered adult female deaths characterized by unusually high mortality of young adults 28,000 relative to older adults. Unfortunately, DHS surveys collect 24,000 sibling histories only from women of reproductive age 20,000 (although respondents report on siblings older than them- selves) and the surveys cover rather small samples for the deaths 16,000 of study of adult mortality, especially at ages 45­59. Thus, these 12,000 surveys lack the statistical power to enable one to produce number 8,000 meaningful estimates of how the age pattern of mortality in 4,000 adulthood is evolving in each country as the overall death 0 rate rises. Timaeus and Jasseh (2004) incorporate the results 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95+ age of more recent DHS surveys into an analysis based on 26 studies in 23 countries. They attempt to sidestep the rela- 1998 1999 2000 2001 2002 2003 tively small size of DHS samples by estimating a common age pattern of mortality increase across all the countries in Source: Bradshaw et al. 2004. which HIV has become present while determining the size of that mortality increase separately for each country. Their in table 4.1 are not consistently exaggerated. For example, smoothed summary estimates of the level and trend in the the estimates for Zimbabwe produced by the WHO are probability of dying between ages 15 and 60 are shown in broadly consistent with the most up-to-date empirical esti- table 4.4 and by the WHO region in figure 4.5, which also mates shown in figure 4.2. plots the WHO's own estimates of the same index for the Although the rise in adult mortality in South Africa by year 2000 (Lopez et al. 2002). 2000 remained moderate at the national level, adult mortality The results of the analysis by Timaeus and Jasseh (2004) had already risen far more in some parts of the country. Data show that the fastest rises in adult mortality have occurred from a demographic surveillance site in a rural area of one of in South Africa, Zimbabwe, Zambia, Uganda, Guinea, and the provinces that has been most severely affected by Cameroon. Adult mortality has risen relatively slowly or HIV/AIDS show that a massive rise in adult mortality continued to fall in the Sahel. A significant change in the occurred between the mid-1990s and 2000. By the latter year trend about four years after the development of a general- the probability of dying between exact ages 15 and 60 reached ized HIV/AIDS epidemic is observed in 16 of the 19 coun- 58 percent for women and 70 percent for men (Hosegood, tries in which the sibling histories cover the relevant period. Vanneste, and Timaeus 2004). Moreover, this site collects Of the three countries in which HIV has become prevalent information on causes of death using verbal autopsies. As but mortality has not risen markedly, the data for Nigeria they are based on a review of symptoms, these data are less are known to be of poor quality, but it is unclear why liable than the official statistics to attribute AIDS deaths to Ethiopia and Rwanda do not conform to the usual pattern. other causes. They show that the huge rise in adult mortality, The increase in mortality is concentrated among women age Levels and Trends of Adult Mortality | 37 Figure 4.5 Trend in the Estimates of the Probabilities of Dying between Ages 15 and 60 (45q15), by Sex and WHO Region, from DHS Sibling Histories (per 1,000 people) Southern Africa--men Southern Africa--women 0.7 0.7 60 60 0.6 0.6 and and 15 15 0.5 0.5 0.4 0.4 between between 0.3 0.3 dying dying of of 0.2 0.2 0.1 0.1 probability probability 0 0 1985 1990 1995 2000 1985 1990 1995 2000 year year Malawi Mozambique South Africa Malawi Mozambique South Africa Zambia Zimbabwe Zambia Zimbabwe Eastern Africa--men Eastern Africa--women 0.7 0.7 60 60 0.6 0.6 and and 15 15 0.5 0.5 0.4 0.4 between between 0.3 0.3 dying dying of of 0.2 0.2 0.1 0.1 probability probability 0 0 1985 1990 1995 2000 1985 1990 1995 2000 year year Ethiopia Kenya Rwanda Ethiopia Kenya Rwanda Uganda Tanzania Uganda Tanzania Source: Timaeus and Jasseh 2004. 25 to 39 and men age 30 to 44. On average, men's death rates women, the discrepancies between the two series are larger have risen by about a third more than those of women. and more widespread. They extend to South Africa, the east- However, as women are dying at younger ages and African ern African countries and, to a lesser extent, Côte d'Ivoire, populations have grown rapidly during the last few decades, Guinea, and Togo in western Africa. These discrepancies the sex differential in the number of AIDS deaths is small. might result from severe underreporting of dead siblings in Extrapolation of the trend in the sibling history estimates the DHS. However, it is unclear why this should be more forward to 2000 and comparison of them with those made serious in lower mortality countries and, in some cases, for by the WHO (also shown in figure 4.5) reveal that they sisters alone. Alternatively, the UN agencies may tend to are in quite good agreement for some countries, including overestimate adult mortality in Africa. An analysis compar- high mortality countries, such as Malawi, Zambia, and ing the numbers of orphans implied in the estimates by the Zimbabwe. For men, the WHO estimates are much higher Population Division with empirical data gathered by the than predicted by the sibling history estimates in DHS on the proportions of children orphaned provides evi- Mozambique and in countries where the prevalence of HIV dence in support of the latter interpretation (Grassly et al. infection is fairly low, that is to say Chad and western 2004). The DHS data find far fewer orphans than are pre- African countries such as Benin, Burkina Faso, and Mali. For dicted by models based on the Population Division and 38 | Debbie Bradshaw and Ian M. Timaeus Figure 4.5 (Continued ) Central Africa--men Central Africa--women 0.7 0.7 60 60 0.6 0.6 and and 15 15 0.5 0.5 0.4 0.4 between between 0.3 0.3 dying dying of of 0.2 0.2 0.1 0.1 probability probability 0 0 1985 1990 1995 2000 1985 1990 1995 2000 year year Cameroon Central African Rep. Chad Cameroon Central African Rep. Chad Western Africa--men Western Africa--women 0.7 0.7 60 60 0.6 0.6 and and 15 15 0.5 0.5 0.4 0.4 between between 0.3 0.3 dying dying of of 0.2 0.2 0.1 0.1 probability probability 0 0 1985 1990 1995 2000 1985 1990 1995 2000 year year Benin Burkina Faso Côte d'Ivoire Benin Burkina Faso Côte d'Ivoire Guinea Mali Togo Guinea Mali Togo UNAIDS estimates of mortality. Just as in the comparison of evidence shows that adult mortality rates are generally high, the sibling history estimates with the WHO estimates reflecting poor levels of health in the region. By the mid- presented here, these discrepancies tend to be larger for 1990s the levels of adult mortality do appear to have become maternal orphans than paternal orphans and larger in low- far more varied among the regions and countries and mortality countries than in high-mortality countries. This between the genders than they had been a decade earlier. pattern of discrepancies suggests that, rather than the prob- Many countries have shown an increase, with death rates lem being underreporting of orphanhood in DHS or exag- reaching unprecedented heights. This is in stark contrast to geration of the scale of AIDS mortality, the model life table the findings by Hill (2003) that adult mortality has been systems used by the UN agencies may systematically overes- declining in most developing countries--at a rate of 1 per- timate background adult mortality in Sub-Saharan Africa, cent per year for men and 2 percent per year for women. especially for women. Estimates of adult mortality for some countries in Sub- Saharan Africa are by far the highest in the world. CONCLUSION Because of the inadequacy of vital statistics in Sub- Saharan Africa, both the UN Population Division and the The lack of good quality vital registration data in Sub- WHO estimate adult mortality and life expectancy in the Saharan Africa has resulted in uncertainty in the levels of countries of the region by extrapolating from estimates of adult mortality. Estimates of adult mortality vary, depend- child mortality using model life tables and adding in an esti- ing on the data source, the methodology used, and the mate of AIDS deaths based on antenatal clinic data. Enough assumptions made. Despite the uncertainties in the data, the is known about adult mortality in Africa to make it clear Levels and Trends of Adult Mortality | 39 Table 4.4 Estimates of the Probability of Dying between Ages 15 and 60(45q15), from DHS Sibling Histories (per person) Women Men Country 1985 1990 1995 2000 1985 1990 1995 2000 Benin 0.201 0.187 0.173 -- 0.250 0.232 0.216 -- Burkina Faso -- 0.225 0.237 0.251 -- 0.271 0.305 0.346 Cameroon -- 0.147 0.226 -- -- 0.179 0.276 -- Central African Republic 0.308 0.367 0.469 -- 0.348 0.418 0.558 -- Chad -- 0.206 0.219 -- -- 0.189 0.207 -- Côte d'Ivoire 0.204 0.252 0.340 -- 0.257 0.319 0.452 -- Ethiopia -- 0.433 0.342 0.266 -- 0.503 0.423 0.359 Guinea -- 0.189 0.180 0.325 -- 0.197 0.188 0.357 Kenya -- 0.175 0.262 -- -- 0.185 0.292 -- Malawi 0.275 0.306 0.435 -- 0.272 0.306 0.464 -- Mali 0.244 0.230 0.219 -- 0.259 0.244 0.236 -- Mozambique -- 0.147 0.181 -- -- 0.181 0.224 -- Namibia 0.172 0.183 -- -- 0.318 0.336 -- -- Niger 0.251 0.198 -- -- 0.252 0.199 -- -- Nigeria -- 0.076 0.114 0.179 -- 0.093 0.140 0.233 Rwanda -- -- 0.458 0.363 -- -- 0.610 0.533 Senegal 0.165 0.163 -- -- 0.186 0.184 -- -- South Africa -- 0.135 0.147 -- -- 0.280 0.302 -- Tanzania 0.127 0.184 0.266 -- 0.166 0.236 0.360 -- Togo -- 0.162 0.204 -- -- 0.187 0.238 -- Uganda 0.274 0.339 0.416 0.415 0.310 0.395 0.511 0.545 Zambia 0.149 0.339 0.547 -- 0.159 0.372 0.623 -- Zimbabwe 0.094 0.187 0.378 0.566 0.113 0.232 0.486 0.726 Source: Timaeus and Jasseh 2004. Notes: -- not available. Estimates extrapolated to 2000 are presented for countries conducting a survey in 1998 or later. that age patterns of mortality vary markedly between differ- estimates of AIDS mortality made by UNAIDS and the ent populations in the region. The assumption that a fixed WHO. Nevertheless, the evidence that adult mortality is relationship between adult and child mortality holds across now rising in most of Sub-Saharan Africa is clear cut. the continent is bound to produce estimates that are badly Moreover, the empirical evidence as to the size and speed of in error in particular countries, even if they are unbiased the rise in adult mortality in different countries in the region overall. In addition, evidence from DHS orphanhood data and the evolving age pattern of the mortality impact is con- (Grassly et al. 2004), DHS sibling histories (Timaeus and sistent with predictions from epidemiological surveillance Jasseh 2004), and earlier surveys and censuses (Timaeus data and models of the impact of HIV/AIDS on adult mor- 1993) all suggest that in recent years the Population tality in most countries. The annual toll of AIDS deaths in Division's extrapolations using Princeton North model life Sub-Saharan Africa continues to grow rapidly, and the tables have overestimated non-AIDS adult mortality in large urgency of responding to the poor health conditions in parts of the region. The WHO's estimates infer the relation- the region has been exacerbated by the emergence of the ship between child and adult mortality from a database of HIV/AIDS epidemic. populations that tend to have lower mortality than is typical In some African countries significant numbers of people of Africa and appear to suffer from very similar limitations have been dying from AIDS since the 1980s. In such coun- to those produced by the Population Division. tries it is difficult to imagine what would have happened to Given the limitations of the demographic and epidemio- rates of mortality by today in the absence of AIDS. Thus, it logical data, there are inevitably large errors in some of the is no longer viable to estimate and forecast mortality by 40 | Debbie Bradshaw and Ian M. Timaeus adding an estimate of AIDS deaths to a separate estimate of invaluable and can yield relatively reliable statistics on non-AIDS mortality. Instead, new approaches for making causes of death. However, it needs to be balanced by other mortality estimates and projections are required that can be sources that can provide nationally representative data. calibrated to data on both the prevalence of HIV infection Although sibling history data have serious limitations, in and all-cause adult mortality. particular their limited statistical precision in all but the While existing data do signal the broad scale of the impact largest surveys, they have proved useful. We believe that of AIDS on the mortality of adult populations in Sub- countries should continue to collect these data. Moreover, Saharan Africa, they are clearly not sufficiently detailed, reli- they should collect them repeatedly, as this greatly increases able, or timely to provide the description of changes in level the scope for assessment of the quality of the statistics that and age pattern that is urgently needed to monitor the result. However, the crucial source of demographic data for impacts of programs that are being funded by African gov- planning, particularly at the subnational or district level, is ernments, the Global Fund, the World Bank, and bilateral the population census. All African governments are faced donors. The fundamental cause of the uncertainty about with either the reality or the threat of massive rises in adult levels of adult mortality and the impact on them of AIDS in mortality and they should all include questions about Africa remains the limitations of epidemiological monitor- household deaths by age and sex and about orphanhood in ing and demographic data systems in the region. The rollout their 2010-round censuses. A major drive is needed on the of antiretroviral care and treatment activities throughout the part of the United Nations agencies and donor organiza- continent should have a major impact (short, middle, and tions to support African governments in this, so as to ensure long term) on mortality dynamics and the demographic that question and form design, enumerator training, data landscape of the region. More than ever before, better epi- capture, and the analysis of these data are all conducted demiological and demographic data are required to evaluate successfully. the impact of health programs and to assess whether these efforts are effective in reducing the mortality of young adults. National statistical offices, technical advisers, donors, and REFERENCES researchers are urged to assign more priority to the collec- Bennett, N., and S. Horiuchi. 1984. "Mortality Estimation from Registered tion of adult mortality data in this region. In addition to Deaths in Less Developed Countries." Demography 21: 217­33. responding to the poor levels of adult health, reinvigorating Bicego, G. 1997. "Estimating Adult Mortality Rates in the Context of the vital registration and statistics is essential in the long term AIDS Epidemic in Sub-Saharan Africa: Analysis of DHS Sibling Histories." Health Transition Review 7 (Suppl. 2): 7­22. and must also be placed high on the health and develop- Blacker, J. 2004. "The Impact of AIDS on Adult Mortality: Evidence from ment agenda. An effective registration system is the only way National and Regional Statistics." AIDS (Suppl. 2): S10­26. of producing detailed annual series of mortality statistics, Blacker, J., P. Kizito, and B. Obonyo. 2003. "Projecting Kenya's Mortality: including those on causes of death. Moreover, as countries Using Spectrum to Project the AIDS and Non-AIDS Components." Paper presented at the Conference on Empirical Evidence for the in Sub-Saharan Africa experience the phenomenon of Demographic and Socioeconomic Impact of AIDS, Durban. aging, improved vital registration systems will be critical to Botha, J. L., and D. Bradshaw. 1985. "African Vital Statistics--A Black provide information about mortality trends at older ages, Hole?" South African Medical Journal 67: 977­81. which are largely unknown at present and difficult to inves- Bradshaw, D., R. Laubscher, R. Dorrington, D. Bourne, and I. M. Timaeus. 2004. "Unabated Rise in Number of Adult Deaths in South Africa." tigate using methods based on survey data. Thus, countries South African Medical Journal 94 (4): 278­79. are encouraged to establish expert teams to critically review Brass, W. 1975. Methods for Estimating Fertility and Mortality from Limited their national civil registration systems in regard to their and Defective Data. Chapel Hill, NC: International Program of Laboratories for Population Statistics. legal framework, organizational issues, system design, train- Brass, W., and K. H. Hill. 1973. "Estimating Adult Mortality from ing needs, and quality control issues and to implement Orphanhood." In International Population Conference, Liège, 1973. strategies for their improvement. Vol. 3. Liège: International Union for the Scientific Study of Given the stresses on governments in the region (in part Population. Coale, A. J., P. Demeny, and B. Vaughan. 1983. Regional Model Life Tables due to the impacts we are trying to measure), it is also vital and Stable Populations. London: Academic Press. to propose ways in which survey data, census data, and Dorrington, R. E., D. Bourne, D. Bradshaw, R. Laubscher, and I. M. demographic surveillance data might be expanded and bet- Timaeus. 2001. The Impact of HIV/AIDS on Adult Mortality in South ter used to track and understand changes in mortality while Africa. Cape Town: Medical Research Council. http://www.mrc/bod/ bod.htm. vital registration still remains defective. Demographic and Dorrington, R., T. A. Moultrie, and I. M. Timaeus. 2004. Estimation of epidemiological surveillance in localized sites has proved Mortality Using the South African Census 2001 Data. Monograph 11. Levels and Trends of Adult Mortality | 41 Cape Town, South Africa: Centre for Actuarial Research, University Demographic and Health Surveys World Conference, Washington, D.C. of Cape Town. http://www.commerce.uct.ac.za/care/Monographs/ Vol. 3. Columbia, MD: IRD/Macro International. Monographs/Mono11.pdf. Salomon, J. A., and C. D. L. Murray. 2001."Modeling HIV/AIDS Epidemics Feeney, G. 2001. "The Impact of HIV/AIDS on Adult Mortality in in Sub-Saharan Africa Using Seroprevalence Data from Antenatal Zimbabwe." Population and Development Review 27: 771­80. Clinics." Bulletin of the World Health Organization 79: 586­692. Grassly, N. C., J. J. C. Lewis, M. Mahy, N. Walker, and I. M. Timaeus. 2004. Sewankambo, N. K., R. H. Gray, S. Ahmad, D. Serwadda, F. Wabwire- "Comparison of Survey Estimates with UNAIDS/WHO Projections of Mangen, F. Nalugoda, N. Kiwanuka, et al. 2000. "Mortality Associated Mortality and Orphan Numbers in Sub-Saharan Africa." Population with HIV Infection in Rural Rakai District, Uganda." AIDS Studies 58: 207­17. 14: 2391­400. Gregson, S., R. M. Anderson, J. Ndlovu, T. Zhuway, and S. K. Chandiwana. Stanton, C., N. Abderrahim, and K. Hill. 1997. DHS Maternal Mortality 1997."Recent Upturn in Mortality in Rural Zimbabwe: Evidence for an Indicators: An Assessment of Data Quality and Implications for Data Use. Early Demographic Impact of HIV-1 Infection?" AIDS 11: 1269­80. Demographic and Health Surveys Analytical Report 4. Calverton, MD: Macro International. Hill, K. 1987. "Estimating Census and Death Registration Completeness." Asian and Pacific Population Forum 1 (3): 8­13, 23­24. Timaeus, I. M. 1991. "Estimation of Mortality from Orphanhood in Adulthood." Demography 28: 213­27. ------. 2003. "Adult Mortality in the Developing World; What We Know and How We Know It." Paper presented at the Training Workshop on ------. 1992. "Estimation of Adult Mortality from Paternal Orphanhood: HIV/AIDS and Adult Mortality in Developing Countries, New York, A Reassessment and a New Approach." Population Bulletin of the September. Organized by the United Nations Population Division, United Nations 33: 47­63. Department of Economic and Social Affairs, UN/POP/MORT/2003/1. ------. 1993. "Adult Mortality." In Demographic Change in Sub-Saharan http://www.un.org/esa/population/publications/adultmort/HILL_ Africa, ed. K. A. Foote, K. H. Hill, and L. G. Martin. Washington, DC: Paper1.pdf. National Academy Press. Hosegood, V., A. M. Vanneste, and I. M. Timaeus. 2004. "Levels and Causes ------. 1998. "Impact of the HIV Epidemic on Mortality in Sub-Saharan of Adult Mortality in Rural South Africa: The Impact of AIDS." AIDS Africa: Evidence from National Surveys and Censuses." AIDS 12 18: 663­71. (Suppl. 1): S15­27. Kitange, H. M., H. Machibya, J. Black, D. M. Mtasiwa, G. Masuki, D. ------. 1999. "Mortality in Sub-Saharan Africa." In Health and Mortality: Whiting, N. Unwin, et al. 1996. "Outlook for Survivors of Childhood Issues of Global Concern, ed. J. Chamie and R. L. Cliquet. New York and in Sub-Saharan Africa: Adult Mortality in Tanzania." British Medical Brussels: United Nations Population Division and Population and Journal 312: 216­20. Family Study Centre, Flemish Scientific Institute. Kowal, P., C. Rao, and C. Mathers. 2003. Report on a WHO Workshop: Timaeus, I. M., R. E. Dorrington, D. Bradshaw, N. Nannan, and D. Bourne. Minimum Dataset on Ageing and Adult Mortality in sub Saharan Africa. 2001. "Adult Mortality in South Africa, 1980­2000: From Apartheid to Geneva: WHO. AIDS." Paper presented at the Population Association of America's Lopez, A. D., O. B. Ahmad, M. Guillot, B. D. Ferguson, J. A. Salomon, C. J. L. annual meeting, Washington, DC, March 29­31. Murray, and K. H. Hill. 2002. World Mortality in 2000: Life Tables for Timaeus, I. M., and M. Jasseh. 2004. "Adult Mortality in Sub-Saharan 191 Countries. Geneva: WHO. Africa: Evidence from Demographic and Health Surveys." Demography Murray, C. J. L., B. D. Ferguson, A. D. Lopez, M. Guillot, J. A. Salomon, and 41: 757­72. O. Ahmad. 2003. "Modified Logit Life Table System: Principles, Timaeus, I. M., and A. J. Nunn. 1997. "Measurement of Adult Mortality in Empirical Validation, and Application." Population Studies 57: 165­82. Populations Affected by AIDS: An Assessment of the Orphanhood Method." Health Transition Review 7 (Suppl. 2): 23­43. Murray, C. J. L., and A. D. Lopez. 1996. The Global Burden of Disease: A Comprehensive Assessment of Mortality and Disability from Diseases, Todd, J., R. Balira, H. Grosskurth, P. Mayaud, F. Mosha, G. ka-Gina, A. Injuries and Risk Factors in 1990 and Projected to 2020. Vol. 1 of the Klokke, et al. 1997. "HIV-Associated Adult Mortality in a Rural Global Burden of Disease and Injury series. Cambridge, MA: Harvard Tanzanian Population." AIDS 11: 801­7. University Press. Tollman, S. M., K. Kahn, M. Garenne, and J. S. Gear. 1999. "Reversal in Murray, C. J. L., G. Yang, and X. Qiao. 1992. "Adult Mortality: Levels, Mortality Trends: Evidence from the Agincourt Field Site, South Africa, Patterns, and Causes." In The Health of Adults in the Developing World, 1992­1995." AIDS 13: 1091­97. ed. R. G. A. Feachem, T. Kjellstrom, C. J. L. Murray, O. Mead, and M. A. UNAIDS (Joint United Nations Programme on HIV/AIDS) Reference Phillips. New York: Oxford University Press. Group on Estimates, Modelling and Projections. 2002. "Improved Notkola, V., I. M. Timaeus, and H. Siiskonen. 2004. "Impact on Mortality Methods and Assumptions for Estimation of the HIV/AIDS Epidemic of the AIDS Epidemic in Northern Namibia Assessed Using Parish and Its Impact: Recommendations of the UNAIDS Reference Group Registers." AIDS 18: 1061­65. on Estimates, Modelling and Projections." AIDS 16: W1­W16. Nunn, A. J., D. W. Mulder, A. Kamali, A. Ruberantwari, J.-F. Kengeya- United Nations. 1982. Model Life Tables for Developing Countries. New Kayondo, and J. Whitworth. 1997. "Mortality Associated with HIV-1 York: United Nations. Infection over Five Years in a Rural Ugandan Population: Cohort ------. 2002. World Population Prospects: The 2000 Revision. Vol. 3: Study." British Medical Journal 315: 767­71. Analytical Report. New York: United Nations. Pison, G., and A. Langaney. 1985. "The Level and Age Pattern of Mortality ------. 2003. World Population Prospects: The 2002 Revision; Datasets in in Bandafassi (Eastern Senegal): Results from a Small-Scale and Excel and PDF Formats. New York: United Nations. Intensive Multi-Round Survey." Population Studies 39: 387­405. Urassa, M., J. T. Boerma, R. Isingo, J. Ngalula, J. Ng'weshemi, G. Mwaluko, Preston, S. H., and N. G. Bennett. 1983. "A Census-Based Method for and B. Zaba. 2001. "The Impact of HIV/AIDS on Mortality and Estimating Adult Mortality." Population Studies 37: 91­104. Household Mobility in Rural Tanzania." AIDS 15: 2017­23. Rutenberg, N., and J. Sullivan. 1991. "Direct and Indirect Estimates of WHO (World Health Organization). 2000. The World Health Report Maternal Mortality from the Sisterhood Method." In Proceedings of the 2000--Health Systems: Improving Performance. Geneva: WHO. 42 | Debbie Bradshaw and Ian M. Timaeus Chapter 5 Causes of Death Chalapati Rao, Alan D. Lopez, and Yusuf Hemed Consistent estimates of cause-specific mortality are essential the expansion in available data suggests that estimates of for understanding the overall epidemiological profile of dis- causes of death can now be made with somewhat greater ease in a population. The principal data source for these esti- confidence. The absence of complete vital registration data mates is civil registration systems. Adequately functioning in virtually all countries of the region nonetheless means that systems that produce statistics on causes of death on a regu- we need to rely on epidemiological research and demo- lar basis exist in only about one-third of all countries of the graphic surveillance to generate model-based estimates of world (Lopez et al. 2002). In Sub-Saharan Africa, very little deaths by cause. information has been available on cause-specific mortality, Such an estimation process is complex and involves two let alone data from civil registration systems, as described in stages. First, a demographic estimate of overall mortality the previous edition of this book (Feachem and Jamison by age and sex is required. Second, a cause-specific mortality 1991). Estimates at that time were derived largely from inde- structure is fitted to this estimate. Many assumptions are pendent disease-specific epidemiological studies and were required, and an attempt has been made here to delineate not examined within the context of an overall demographic these clearly, so that they can be kept in mind when inter- "envelope" of mortality, as is required to ensure that claims preting the results. about causes of death are not exaggerated. Over the past decade, much progress has been made in the collection of mortality statistics from a wide array of sources. DATA SOURCES These include data from previously existing sources that were uncovered during a systematic search, as well as data The previous edition of this book provided information on from new data collection ventures that were established to fill cause-specific mortality in the form of a set of estimates for these data gaps. Although we are still a long way from having the entire region, disaggregated over seven broad disease satisfactory empirical data that can be directly used to derive categories, and in six age groups. These estimates were national and regional cause-specific mortality estimates, essentially derived from linear models relating mortality 43 levels to broad causes of death (Preston 1976), with very lit- Table 5.1 Proportion of Deaths Due to Diseases and Injuries tle attempt to incorporate other epidemiological informa- in Children under Five Years, by Source of Data tion. The authors stated that the large proportion of deaths No. of No. of Gp 1 Gp 2 Gp 3 Othersa lumped under other causes reflected the weakness of the Method studies deaths (%) (%) (%) (%) Total estimates, and suggested great caution in interpreting or Verbal autopsy 13 4,964 68 4 0 28 100 using them (Feachem, Jamison, and Bos 1991). (surveys) Building on previous experience, researchers have made Verbal autopsy and 3 1,139 54 16 0 30 100 much more comprehensive attempts over the past decade to hospital records collect and analyze information in order to estimate the Verbal autopsy 4 1,906 62 1 0 37 100 (prospective studies) cause-specific mortality structure for Sub-Saharan Africa Vital registration 3 11,878 52 19 0 29 100 as part of the Global Burden of Disease (GBD) Study Hospital records 1 2,419 75 0 0 25 100 (Mathers, Stein et al. 2002; Murray and Lopez 1996). Other multiple 3 1,288 86 9 1 4 100 Discussed here are the results of two major data collection sources efforts that were undertaken to estimate the burden of pre- Hospital autopsy 1 953 73 15 9 12 100 mature mortality in Africa. The first effort involved a search for, and analysis of, epidemiological literature on causes of Source: Adetunji, Murray, and Evans 1996. a. Includes ill-defined conditions and cause unknown. death in the region. The second effort involved a similar exercise using data from national vital records. The high proportion of deaths for which the cause was unknown raises serious questions about data quality. The Review of Epidemiological Literature largest number of deaths in this review was in the vital Much of the earlier attempts to estimate causes of death in records for Abidjan, Côte d'Ivoire (1987­92). However, the region drew on published and unpublished reports these too have high proportions of "other" causes (29 per- based on epidemiological studies. Adetunji, Murray, and cent). A detailed analysis of specific causes based on the Evans (1996) conducted a major review of epidemiological data from these studies suggested that respiratory infections, studies to identify usable cause-specific mortality data. They diarrhea, malnutrition, and anemia were the leading causes identified a total of 48 research studies from 16 countries of death in this age group. that met specific criteria related to data definition and data At ages 5 to 14 years (see table 5.2), the largest number of quality issues for inclusion in the analysis. The main criteria deaths was also in the vital records from the same source, for inclusion was that the study should have reported the although the total number of deaths registered is very small. relative magnitude of all causes of death and not focus on a In such situations, deriving a meaningful estimate of even single or a few causes without mention of other causes with- the proportions across broad cause groups is difficult, in the sample. Other criteria refer to methods used to derive because paucity in numbers is coupled with high propor- the cause of death, age groups reported, site, and period of tions of ill-defined causes (44 percent). Nevertheless, the the study. Reports from South Africa were excluded from results suggested that malaria, diarrheal diseases, and this analysis, since the data overlapped with that available malnutrition were the leading causes of death among from vital records. About 80 percent of the reports were school-age children. based on information from hospital records, verbal autopsy interviews, or both. A majority of the studies (again 80 per- Table 5.2 Proportion of Deaths Due to Diseases and Injuries cent) focused on causes of maternal and child mortality, in Children Age 5 to 14 Years, by Source of Data whereas causes of adult male deaths were relatively neg- No. of No. of Gp 1 Gp 2 Gp 3 Othersa lected. The authors presented results in the form of cause Method studies deaths (%) (%) (%) (%) Total composition of mortality for three distinct cause groups: Verbal autopsy 1 25 88 0 0 12 100 communicable diseases and maternal, perinatal, and nutri- (surveys) tional conditions (group 1); noncommunicable diseases Vital registration 2 2,817 38 12 6 44 100 (group 2); and injuries (group 3). These cause groups were Hospital records 1 713 67 0 21 12 100 the same as those used in the GBD Study. The results of their Hospital autopsy 1 52 56 15 12 17 100 review of deaths of children younger than five years are Source: Adetunji, Murray, and Evans 1996. summarized in table 5.1. a. Includes ill-defined conditions and cause unknown. 44 | Chalapati Rao, Alan D. Lopez, and Yusuf Hemed For ages above 15 years, epidemiological studies on and such epidemiological estimates for individual causes causes of death concentrate on causes of maternal mortality. are useful to build the overall cause-specific mortality Adetunji, Murray, and Evans (1996) reviewed information picture. from 21 studies and identified 3,818 deaths due to maternal causes. About 80 percent of these studies were based on Information from Vital Records hospital records. This could have two possible implications. On the one hand, hospital-based studies could overesti- Another data collection effort was centered on the acquisi- mate maternal mortality, as high risk and emergency cases tion of vital records data. The World Health Organization are likely to be concentrated in hospitals. On the other (WHO) conducted a comprehensive search for these data hand, better emergency care in hospitals may avert more over the period 2001­02, as part of the data collection for deaths than would have occurred in the community. the GBD 2000 project (Kowal, Rao, and Mathers 2003). Few Notwithstanding these possible biases, the causes as docu- countries in Africa have vital registration systems that are mented in hospital records are likely to be more reliable than more than 50 percent complete. Coverage is about this level those from community-based studies. The pooled results in Kenya and Zimbabwe (Lopez et al. 2002), and close to suggested hemorrhage (19 percent), puerperal sepsis 90 percent in South Africa (Dorrington et al. 2001). In (13 percent), hypertensive disorders of pregnancy (7.8 per- Mozambique a major exercise was undertaken to improve cent), and ruptured uterus (7 percent) as the leading causes mortality registration and cause-of-death attribution of maternal mortality in Sub-Saharan Africa, at least among in four cities (Cliff et al. 2003). The ministries of health in women referred to hospitals. Botswana, Eritrea, and Zambia collect and collate informa- Adetunji, Murray, and Evans (1996) also documented tion on causes of death from health facility data and from three sources of data on causes of adult mortality; two from vital records, essentially in urban areas. The AMMP in vital records in Lagos (1977) and Abidjan (1987­92), and Tanzania has built up a comprehensive district-level mor- one from the Adult Morbidity and Mortality Project tality surveillance system that operates currently in three (AMMP) in Tanzania (1992). The data from Lagos predate districts (Hai, Morogoro, and Dar es Salaam) and has the HIV/AIDS period and record accidents and violence compiled information on causes of death over a 10-year as the leading cause of adult deaths (26.3 percent). In period. Vital records data from the city of Antananarivo have Abidjan, the data suggest hypertensive disease (31 percent), been systematically compiled over a 12-year period to diarrheal disease (11 percent) and HIV/AIDS (10.5 percent) describe urban causes of death in Madagascar (Waltisperger, as the leading causes. Data from the AMMP are available by Cantrelle, and Ralijaona 1998). Data from all these countries sex and show HIV/AIDS to be the leading cause in both have been collated and analyzed at the WHO. sexes (higher proportions in women) and injuries in males Vital records data, although impressive in overall num- and pregnancy-related causes to be the other significant bers of deaths for which information on cause is available, conditions. All these results suggest that hypertensive dis- have their own limitations. First, coverage is incomplete, and ease, HIV/AIDS, pregnancy-related causes, and injuries are it is not possible to map these data to an underlying popu- the leading causes of death among adults in Sub-Saharan lation to make an assessment of death rates. Second, there is Africa. no indication of the bases used for cause-of-death attribu- We have described this review in some detail in part tion, whether physician attribution at the time of death, because it was the most comprehensive previous attempt to hospital records, verbal autopsy, or lay reported informa- estimate causes of death in Africa, and in part to show that tion. The proportionate distribution of causes of death are compiling information from various sources, despite the therefore examined in this chapter according to the GBD enormous effort involved, still results in substantial groups at different ages, as was done in the earlier analyses uncertainty about the cause structure of mortality owing to using data from epidemiological studies. biases in the way the data were collected and the high Included in these tabulations are the proportionate proportions of unspecified causes in the reports. However, distributions for two epidemiological subregions in Sub- these epidemiological studies were designed to focus on one Saharan Africa (AFR D and AFR E),1 as estimated by the or a few specific causes of mortality, yielding useful GBD cause-of-death study. These subregions group coun- information on incidence, duration, and case-fatality rates. tries with similar estimated levels of child and adult These indexes were used by independent researchers to infer mortality, as described in detail later in the section on the mortality specific to that cause at the population level, GBD estimation process. For comparison purposes, and so Causes of Death | 45 Table 5.3 National Vital Records Data: Proportionate Table 5.4 National Vital Records Data: Proportionate Distribution of Cause of Death of Children under Five Distribution of Cause of Death at Age 5 to 14 Years Year of data No. of Gp 1 Gp 2 Gp 3 Ill-defineda Year of data No. of Gp 1 Gp 2 Gp 3 Ill-defineda Country collection deaths (%) (%) (%) (%) Country collection deaths (%) (%) (%) (%) Botswana 1998 1,443 80 10 1 9 Botswana 1998 129 53 30 9 4 Kenya 2001 53,478 88 4 1 7 Kenya 2001 11,240 75 9 5 12 Madagascar 1984­95 49,250 83 7 2 8 Madagascar 1984­95 5,362 57 19 12 12 Mozambique 2001 6,712 78 6 3 13 Mozambique 2001 1,013 68 12 13 7 South Africa 1996 29,043 75 8 8 9 South Africa 1996 5,160 19 19 53 10 Tanzaniab 1998­2000 1,338 88 2 0 9 Tanzania 1998­2000 653 73 8 12 6 Zambia 1999­2000 32,586 94 2 2 2 Zimbabwe 1995 1,842 47 22 25 6 Zimbabwe 1995 17,786 86 7 3 5 GBD AFR D 2000 248,012 67 6 27 -- GBD AFR D 2000 2,134,635 96 2 2 -- GBD AFR E 2000 314,283 74 6 20 -- GBD AFR E 2000 2,265,960 95 3 2 -- Source: Authors. Note: -- not available. Source: Authors. a. As part of the GBD approach, deaths from ill-defined categories are reallocated to specific Note: -- not available. categories. a. As part of the GBD approach, deaths from ill-defined categories are reallocated to specific categories. b. For Tanzania, information collected in the AMMP project was used. proportions of the groups. Again, from all the data sets, the their plausibility may be judged, these figures are included proportion of deaths due to ill-defined causes is low. In the here with available local data. GBD estimation process, these deaths due to ill-defined Among children under five, the highest proportion of causes at ages above five years are reallocated proportion- deaths, as expected, is from group 1 causes (perinatal condi- ately across group 2 causes. In comparison, group 1 propor- tions, communicable diseases, and malnutrition), ranging tions in the estimates again approximate the higher end of from 75 percent in the South African data set to 94 percent the range of observations in the local data. in the data from Zambia. A reassuring feature of these vital Much more complex is to estimate cause-specific records data is the relatively low proportion of deaths not mortality for the young adult (15 to 44 years) age group. At classified in any of the three cause groups. In the GBD these ages in Sub-Saharan Africa, a triple burden of cause- estimates, unclassifiable deaths at ages below five years specific factors exists, namely, the HIV/AIDS epidemic; high are reallocated to group 1 (Murray and Lopez 1996). incidence rates of injury and violence, especially among Hence, we estimate that about 95 percent of deaths at these males; and the high burden of maternal mortality among ages are due to group 1 causes, which approximates the females. These aspects of cause-specific mortality are upper end of the range of observed proportions from this evident from the results of the epidemiological literature cause category from the vital records data sets (table 5.3). review discussed earlier. Table 5.5 shows the group-specific As age increases, the proportion of deaths due to proportions from vital records at these ages. The proportion noncommunicable diseases and injuries rises. In global of deaths in vital records due to injuries at these ages comparisons, this shift in cause composition across ages is appears to be low in most countries, given the wars and less evident in Sub-Saharan Africa than in other regions, due violence in the region. The GBD estimates of war-related to the catastrophic HIV/AIDS epidemic. At ages 5 to 14 deaths are included in deaths due to violence and marginally years, the South African data set suggests a low proportion increase the proportions of deaths due to injuries. Also, the of deaths due to group 1 causes, which may be because the relatively low proportion of deaths due to group 1 causes population covered by registration is urban, with higher in the Madagascar data set could be because the population socioeconomic status (expressed as GDP per capita) than covered was urban, or that HIV prevalence, transmission, the national average (table 5.4). Another possible cause for and mortality is low in Madagascar (Ravaoarimalala et al. low group 1 proportions is that the coverage of vital regis- 1998). Group 1 conditions appear to cause a major propor- tration in South Africa is only 90 percent, and the missed tion of mortality at these ages. deaths could be group 1 deaths. However, such corrections Local data for deaths at older ages (45 years and older) from would result only in marginal changes in the specific all sites have significantly higher proportions of unclassifiable 46 | Chalapati Rao, Alan D. Lopez, and Yusuf Hemed Table 5.5 National Vital Records Data: Proportionate Table 5.6 National Vital Records Data: Proportionate Distribution of Cause of Death at Age 15 to 44 Years Distribution of Cause of Death at 45 Years of Age and Older Year of data No. of Gp 1 Gp 2 Gp 3 Ill-defineda Year of data No. of Gp 1 Gp 2 Gp 3 Ill-defineda Country collection deaths (%) (%) (%) (%) Country collection deaths (%) (%) (%) (%) Botswana 1998 2,658 76 19 3 2 Botswana 1998 2,082 44 48 2 4 Kenya 2001 69,910 75 11 6 8 Kenya 2001 64,729 57 27 3 13 Madagascar 1984­95 23,056 39 40 13 9 Madagascar 1984­95 45,166 18 63 3 16 Mozambique 2001 8,008 71 12 9 8 Mozambique 2001 5,843 45 30 5 20 South Africa 1996 83,482 25 22 43 10 South Africa 1996 167,828 13 64 7 15 Tanzania 1998­2000 3,943 72 11 9 9 Tanzania 1998­2000 4,811 48 32 4 16 Zimbabwe 1995 30,124 59 13 8 19 Zimbabwe 1995 28,173 30 35 5 31 GBD AFR D 2000 913,976 69 13 18 -- GBD AFR D 2000 1,337,672 28 67 5 -- GBD AFR E 2000 1,879,690 82 7 11 -- GBD AFR E 2000 1,683,816 37 58 5 -- Source: Authors. Source: Authors. Note: -- not available. Note: -- not available. a. As part of the GBD approach, deaths from ill-defined categories are reallocated to specific a. As part of the GBD approach, deaths from ill-defined categories are reallocated to specific categories. categories. deaths, approaching 31 percent in Zimbabwe (table 5.6). At life table­based age-specific risk of mortality is applied to these ages, one would normally expect higher proportions of a national age-sex population estimate. The purpose of deaths from noncommunicable diseases, which is reflected in this step is to provide an upper limit or "demographic the data. There is a possibility of misclassification between envelope," to constrain cause-of-death estimates within group 2 and group 1, since in some instances individuals the bounds of demographic plausibility. with long-standing noncommunicable diseases develop 2. The second step is to use cause-of-death models to pre- infectious complications during terminal stages, which tend dict a GBD group-specific (broad causes) proportionate to be more readily identified and classified as the underlying composition of mortality for each age-sex group and to cause of death. The estimates are close to the upper end of the apply the proportions to the mortality envelope derived range of proportions of group 2 conditions from local data. in step 1 to derive an age-sex GBD group-specific enve- From the above analyses, vital registration data would lope of mortality. seem to be useful in understanding general patterns of cause- 3. The third step is to fit a condition-specific cause structure specific mortality at broad-cause-group level. However, of mortality for each broad-cause group onto their biases introduced by incomplete or selective coverage and respective envelopes as derived in step 2 to derive esti- difficulty in identifying specific causes beyond the broad- mates of deaths from specific conditions by age and sex cause-group level preclude their usage directly in the estima- for the population. tion process. In situations of incomplete data, the process of estimating mortality from specific causes entails the synthesis In the current analysis, separate estimates were developed of information from different data sources or, if necessary, by for each country in Sub-Saharan Africa, and they were drawing on other regional or international cause-specific summed up to generate two subregional population aggre- mortality patterns. gations. The classification of WHO member states into the mortality strata was carried out using population estimates GBD Process for Estimating Cause-Specific Mortality for 1999 (United Nations Population Division 1998) and The GBD Study method of estimating cause-specific mor- estimates of child mortality (defined as q 5 0 ; the risk of dying tality in populations without detailed information on the between birth and age 5) and adult mortality (defined as levels or cause structure of mortality is essentially based on 45 15 q ; the risk of dying between ages 15 and 60) based on three sequential steps: WHO analyses of mortality rates for 1999 (Mathers, Stein et al. 2002). Five mortality strata were defined in terms of 1. The first step is to derive an overall envelope of mortality, quintiles of the distribution of 5 0 q and 45 15 q (both sexes in terms of estimated numbers of deaths, for each age-sex combined). Adult mortality 45 15 q was regressed on q and 5 0 group within the population. In practical terms, a model the regression line used to divide countries with high child Causes of Death | 47 Figure 5.1 Global Mortality Strata for GBD 2000 Regions age 15 to 49 years. In a few countries, information on sibling survival, subsequently analyzed to derive estimates of adult 0.8 mortality, has also been collected and compiled. C E A new modified logit model life table system (Murray 0.6 et al. 2003) allows for predicting an abridged life table using as inputs true or estimated levels of child and adult mortal- 15 q 0.4 45 ity or, if necessary, using only levels of child mortality. Recent work by Murray and colleagues has resulted in devel- 0.2 opment of a validated life table system that uses a single global standard that represents the full range of mortality A B D patterns seen in contemporary human populations; the sys- 0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 q tem generates better predictions of age-specific mortality 5 0 rates than the Coale-Demeny or original Brass systems.2 A - Very low child, very low adult B - Low child, low adult C - Low child, high adult D - High child, high adult E - High child, very high adult A set of 1,802 life tables, based on empirical data from national mortality registration systems between 1901 and Source: Mathers, Stein et al. 2002. 1999, were used to develop and test this model. These life tables were based on observed mortality risks in HIV-free mortality into high adult mortality (stratum D) and very populations, and hence, predicted life tables from the model high adult mortality (stratum E), as in figure 5.1. Stratum E can be used to estimate HIV-free age-specific mortality. In includes the countries in Sub-Saharan Africa where this context, estimates of child mortality from DHS surveys HIV/AIDS has had a substantial impact. for the 52 countries in Sub-Saharan Africa were used to A fundamental principle in the estimation process is to develop HIV-free envelopes of mortality at the national maintain internal consistency between the overall demo- level. Steps 2 and 3 as described earlier were carried out for graphic envelope of mortality and the cause composition each country to construct a cause-specific mortality within the envelope. In other words, the mortality due to estimate without accounting for HIV, and as a final step, specific causes within an age group should exactly sum up to epidemiological estimates of HIV mortality were added on the life table­derived envelope of mortality. Such con- to derive the overall national age-, sex-, and cause-specific straints call for significant assumptions and judgments; mortality estimate (Lopez et al. 2002). Figure 5.2 shows, by hence, the final estimates should be interpreted based on way of example, the final age-specific death rates for these choices. With the high estimated levels of HIV/AIDS Zambia, for the year 2000. Note the markedly high death mortality in the region, judgments regarding burden due to rates at young adult ages, including the deaths due to AIDS. competing causes of death are based on evidence from vital Compositional cause-of-death models have been devel- records and epidemiological estimates for specific diseases. oped around the relationship between cause-specific However, preceding these issues is the process of deriving mortality and total mortality by age, as observed from pat- the life tables for estimating the mortality envelope by age terns of causes of death recorded in nations with good vital and sex and the use of cause-of-death models for allocating registration systems (Salomon and Murray 2002). For each deaths to broad-cause groups. age-sex group an income variable for each observed country-year of data has been added to the model to Life Tables strengthen the predicted proportionate composition by the Detailed age-specific mortality rates are required to con- three GBD cause groups. The models have been built using struct life tables for populations. Such rates are not available an empirical data set of 1,576 country-years of observations for any country within Sub-Saharan Africa. Attempts at from 58 countries for the years 1950 to 1998. A set of regres- estimating death rates based on recall of deaths in censuses sion equations have been developed that predict the cause and surveys have generally led to implausibly low death rates composition of mortality by age, for a given input of total (Brass 1968). The best available data on mortality risk are lev- mortality by age and an estimate of national per capita gross els of child mortality derived from Demographic and Health domestic product (GDP). For each country, the estimate of Surveys (DHS) conducted in most countries within the total mortality by age and sex derived from the life tables region during the 1990s. These estimates are derived using (from step 1) and the national GDP are used as inputs, and data from birth histories collected from women respondents the model produces an implied cause composition of 48 | Chalapati Rao, Alan D. Lopez, and Yusuf Hemed Figure 5.2 Estimated Age-Specific Death Rates for Zambia, Figure 5.3 Model-Based Predictions of GBD Cause­Group 2000 Composition of Mortality by Age and Sex, Zambia, 2000 Males 1.000 Males 1.0 rate 0.8 0.100 death 0.6 0.010 age-specific proportion 0.4 log 0.2 0.001 0 15 45 75 age 0 15 30 45 60 75 age Females 1.000 Females 1.0 rate 0.8 0.100 death 0.6 0.010 age-specific proportion 0.4 log 0.2 0.001 0 15 45 75 age 0 15 30 45 60 75 age ASDR (LT) ASDR (with AIDS) GP 1 GP 2 GP 3 Source: Authors. Note: ASDR age-specific death rate; LT from model life tables for Zambia, 2000; graphs are Source: Authors. in logarithmic scale. mortality by age and sex. Since the empirical data set for Epidemiological Estimates developing the regression equations does not include any Cause-of-death models in combination with life table­ information on mortality in countries experiencing AIDS derived mortality risks and national population estimates epidemics, it is difficult to predict with confidence the pro- provide an estimate of the numbers of deaths by age, sex, portions of deaths due to group 1 causes at different ages. and broad-cause group. Estimating the numbers of deaths Wherever possible, the outputs of the model were validated due to specific causes within each broad group--for exam- against compositional structures from available vital records ple, the composition within group 1 due to different infec- data from the country or region and from small-scale longi- tious diseases and perinatal disorders at infant ages or the tudinal population surveillance systems. An example of the composition within group 2 due to neoplasms, cardiovas- model outputs is shown in figure 5.3, depicting the predicted cular diseases, and other noncommunicable diseases at proportionate composition by age for males and females, adult ages--is the next stage. The mass of epidemiological respectively. evidence on mortality due to different infectious diseases Communicable diseases (group 1) account for a large and maternal causes over the past two decades has been proportion of deaths at all ages up to about age 60. The invaluable in this estimation process. Researchers have crit- proportions derived from the model are used to derive GBD ically examined data from different studies and used them group-specific numbers of deaths for each age group from to derive independent epidemiological estimates of cause- the life table­derived total numbers of deaths in each age specific mortality for the region. It is useful to examine in group. After this, information from vital records or specific brief the results of these independent estimation exercises disease epidemiological studies are used to construct the and to assess their plausibility with reference to an overall detailed cause-of-death estimates. Causes of Death | 49 demographic envelope of mortality. We have included in complex interactions between intrinsic factors, such as this discussion results from such estimation exercises avail- nutritional status, and extrinsic factors, such as exposure to able for Africa in 2003. There is a need to conduct similar multiple infections, make the attribution of death to a single estimations for other important conditions, such as tuber- underlying cause extremely difficult. Kosek, Bern, and culosis and road traffic accidents. Guerrant (2003) reviewed 34 studies in 21 countries to derive an estimate of diarrheal mortality rates. Studies were included only if diarrheal deaths were ascertained through Malaria. Malaria is one disease that has been extensively active surveillance, and in the selected studies only a primary researched in Africa. Snow and his colleagues initially esti- cause listed as diarrhea was considered as a diarrheal death. mated malaria mortality in the region for 1990 and later As a result, they estimated a mortality rate of 4.9 per 1,000 refined their approach to produce another set of estimates children under five per year (95 percent confidence interval for 1995 (Snow et al. 1999). They integrated evidence from (CI), 1.0­9.1) in countries with high levels of overall child research on several aspects of the epidemiology of malaria. mortality. On a global scale, this appears as a decline from an First, they used a detailed analysis of environmental factors estimated rate of 13.6 per 1,000 in 1982 (Snyder and Merson that affect the distribution, seasonality, and transmission 1982). intensity to develop an epidemiological stratification of the In view of different cause-of-death attribution strategies continent into five regions based on climatic suitability for in different settings, proportionate mortality due to diar- the existence and stability of malarial transmission. In the rhea is considered as an alternate approach to estimating the next step, they used geographical information systems to number of deaths due to this cause. Morris, Black, and define at-risk populations within regions suitable for trans- Tomaskovic (2003) developed a prediction model to esti- mission based on different transmission scenarios as mate the distribution of deaths among children under five described. A specially constructed population database for by cause. The model estimated that about 20 percent of all more than 4,000 administrative units in Africa (Deichmann under-five deaths in Sub-Saharan Africa would be caused by 1996) was used in this step. Finally, direct estimates of fatal diarrhea. risks for malarial mortality were selected from over 200 Table 5.8 shows the total population of children under empirical data sources on the health impact of malaria, and age five in the region, and the estimate of diarrheal deaths these were combined with estimates of at-risk populations on applying the median observed death rate of 4.9 per 1,000 to derive age-specific estimates of deaths due to malaria in (Kosek, Bern, and Guerrant 2003), and a proportionate the region, as summarized in table 5.7. The median esti- mortality of 20 percent (Black, Morris, and Bryce 2003). mates were taken as the guideline for the GBD estimates for the entire region. Vaccine-Preventable Diseases. Despite expansion in immunization services in the developing world, vaccine- Diarrheal Diseases. Estimates of death rates due to diar- rheal disease from epidemiological studies vary widely across populations, and within populations over time. Also, Table 5.8 Estimates of Diarrheal Deaths of Children under Five, 2000 Estimate based on death rates Table 5.7 Epidemiological Estimates of Malaria Mortality in Under-five population in Sub-Saharan Africa 127,856,836 Nonpregnant African Population, 1995 Estimated diarrheal death rate 4.9 per 1,000 Estimated diarrheal deaths 639,284 Estimated deaths Estimate based on proportionate mortality Age group Lower bound Median Upper bound Under-five total deaths 4,475,675 0 to 4 578,214 765,775 1,010,337 Estimated diarrheal proportionate mortality 20% 5 to 9 109,986 145,747 192,233 Estimated diarrheal deaths 895,135 10 to 14 34,427 45,349 60,215 GBD estimate 626,734 15 22,018 32,730 43,444 Total 744,645 986,601 1,306,229 Source: Estimate based on death rates adapted from Kosek, Bern, and Guerrant 2003; estimate based on proportional mortality from Black, Morris, and Bryce 2003; GBD estimate from Source: Adapted from Snow et al. 1999. WHO 2003. 50 | Chalapati Rao, Alan D. Lopez, and Yusuf Hemed preventable diseases remain significant causes of mortality in Maternal Mortality. Hill, AbouZahr, and Wardlaw (2001) Africa. Stein and colleagues (2003) conducted an epidemio- in collaboration with researchers at the WHO collated avail- logical exercise to estimate measles mortality. They used able evidence on country-specific maternal mortality in the national-level information on vaccine coverage from health form of vital registration records, DHS-type sibling survival surveys and estimates of disease incidence, vaccine efficacy, surveys, and special Reproductive Age Mortality Studies to and case fatality from specific epidemiological studies to develop a statistical model to predict the proportion of develop a static model by which to estimate regional and deaths of women of reproductive ages due to maternal global measles mortality. They estimated a total of 452,000 causes (PMDF). The PMDF was found superior to the deaths due to measles in Sub-Saharan Africa in the year 2000, maternal mortality ratio (MMR) as the dependent variable of which approximately three-fourths would occur among for estimating the number of deaths from maternal causes, children under five. mainly because of its applicability to a demographic "enve- Crowcroft and colleagues (2003) developed a model by lope" of deaths at maternal ages. Also, information from which to estimate cases and deaths due to pertussis for the national-level sisterhood surveys were found to yield more year 1999. Parameters used in the model included vaccine robust measures of PMDF than MMR to use as inputs in the coverage and efficacy data and estimates of case-fatality construction of the model. Independent variables chosen to ratios from epidemiological studies. Two coverage scenarios predict national PMDF were general fertility rate, female lit- were used in the model, at levels below and above 70 per- eracy, per capita income, percentage of deliveries attended cent, based on information from WHO reports adjusted by by skilled attendants, country-specific estimates of HIV survey data where available. For each scenario, a different prevalence, and variables for region and level of vital regis- age structure was used to represent a range of possible tration in individual countries. The predicted PMDF for patterns of infection in susceptible children. The model each country was then applied to age- and sex-specific mor- assumed a vaccine efficacy of 95 percent for preventing tality estimates from World Population Prospects (United death. They estimated 170,000 deaths due to whooping Nations Population Division 1998), to derive the absolute cough in the African region in 1999. numbers of deaths due to maternal causes. Based on the model, Hill, AbouZahr, and Wardlaw (2001) predicted a Lower Respiratory Infections. Acute respiratory infec- total of 272,500 maternal deaths in Sub-Saharan Africa for tions (ARI) are the third leading cause of death globally 1995, with a mean MMR of 1,006 per 100,000 live births. among children under five years of age. Much uncertainty surrounds the ascertainment of pneumonia as the underly- HIV/AIDS. The Joint United Nations Programme on ing cause of death, principally from verbal autopsy­based HIV/AIDS (UNAIDS) and the WHO established an studies in developing countries. In these settings, signifi- Epidemiology Reference Group to work toward making cant comorbidity exists in the form of measles, whooping estimates and projections of mortality on a biennial basis. cough, diarrheal diseases, or malaria. Williams and col- Country-level information on prevalence of infection leagues (2002) developed a model to predict the propor- among attendants at antenatal clinics and sexually transmit- tion of ARI mortality for a given level of under-five mor- ted disease clinics and prevalence among other high-risk tality. A review of 49 studies provided data on the level of groups, such as intravenous drug users, homosexual males, child mortality and the proportion of deaths due to ARI. and commercial sex workers, has been used to monitor epi- The researchers fitted a log linear curve to the data; from demics at the country level. Using these data, an epidemio- the resultant equation they estimated the number of deaths logical model was developed that included the following due to ARI for all countries, using WHO estimates of parameters: country-specific under-five mortality in the year 2000. Through this model they estimated a total of 794,000 1. the initial rate (r) of spread of HIV as determined by the deaths due to ARI in the region in the year 2000. The reproductive potential authors demonstrated that differences between predicted 2. the peak prevalence (f ) as determined by the fraction of proportions from the model and observed proportions population at risk of infection from verbal autopsy studies could be explained by the vari- 3. the final epidemic prevalence ( ) as determined by the ability inherent with the use of verbal autopsy methods, behavioral response of the population induced by such comorbidity. 4. the start date of epidemics in individual countries. Causes of Death | 51 to follow up. The estimated cancer survival rates by site, age, A negative value of indicates that people become less likely and sex for different regions in the world were used as key to adopt risky behavior in response to observed AIDS mor- inputs to estimate the distribution of cancer deaths by site. tality or prevention programs. Hence, apart from prevalence Estimates of cancer mortality by region were then developed data from sentinel sites, behavioral surveys are essential to (Shibuya et al. 2002), correcting for levels of overall mortal- assess the potential of HIV epidemics and to estimate the ity in regions with incomplete coverage of registration, and size of at-risk populations. also correcting for the likely differences in cause-of-death An assessment of factors that affect survival of adults patterns that would be expected in uncovered and often suggested that an overall median survival time of 9 years, poorer subpopulations. As a result, the GBD 2000 Study with a range of uncertainty of 8 to 11 years, and a Weibull estimated a total of 572,000 deaths due to cancer within the distribution of the survival function were used in the region. modeling process. For children, the survival curve was built to account for two periods of high mortality, which are War. Deaths due to war within Sub-Saharan Africa merit infancy, when HIV frequently overwhelms the immature attention as an avoidable burden. Murray and colleagues immune system, and after age nine years, when the response (2002) used a comprehensive analysis of media reports, vital to HIV infection resembles that in adults. Overall, the sur- registration records, and adjustments based on observed vival curve for children predicts 40 percent survival from relationships between direct and indirect mortality to esti- HIV-related mortality at five years of age. mate the global burden of mortality due to armed conflict. The UNAIDS/WHO model was used to develop country- Keeping in mind the limitations of estimates based largely specific point estimates of HIV/AIDS mortality for 2001 on qualitative analysis of media reports, conservative esti- (WHO and UNAIDS 2002). The estimated regional death mates were used for some of the major conflicts in the toll from this disease for Sub-Saharan Africa stands at a total world. According to these estimates, armed conflicts had of 2.2 million deaths. Policy decisions aimed at addressing resulted in an overall death toll of about 77,000 for Sub- this epidemic should include activities aimed at improving Saharan Africa for the year 2001. such measurements. Cancers. Ferlay and colleagues (2001) at the International Estimating Cause-Specific Mortality Agency for Research in Cancer developed a data set of worldwide estimates of cancer incidence, mortality, and Using all the various dimensions of cause-specific mortality prevalence for the year 2000, which they called the Globocan estimation for Sub-Saharan Africa--available data from 2000. Their mortality estimates were based on vital registra- vital records, specific studies, and global disease epidemio- tion data where available; for other regions they used logical extrapolations for the region--we have prepared information from survival models derived from available estimates of causes of death in Sub-Saharan Africa based on cancer registry data (Sankaranarayan, Black, and Parkin the GBD approach. Figure 5.4 summarizes the estimation 1998). These mortality estimates did not correct for under- process for countries within the region. reporting in vital registration or for possible misclassifica- As mentioned earlier, predicted proportions for broad- tion of causes of death. cause groups by age and sex were validated against informa- As part of the GBD 2000 Study, Mathers, Shibuya et al. tion from country-specific vital records where available and (2002) used relative interval survival data from the for neighboring or epidemiologically similar countries Surveillance Epidemiology and End Results (SEER) where possible. Also, wherever applicable, cause-specific Program at the National Cancer Institute in the United proportions from vital records for specific causes were used States to develop an age-period-cohort survival model, in deriving population-level estimates for these causes. The which was further adjusted for levels of economic develop- resultant cause-specific structure was finally adjusted with ment. The model was then applied to age-, sex-, and disease-specific mortality estimates to produce the overall site-specific incidence estimates from cancer registries that numbers of deaths by age, sex, and cause. National estimates were compiled for the Globocan 2000 and from other were summed up to totals for two subregions and for Sub- specific incidence studies. The model was useful to smooth Saharan Africa as a whole. During this process of synthesiz- out random variations in observed incidence and survival ing epidemiological estimates into the overall demographic rates that resulted from small numbers of cases or cases lost and cause-of-death model based on the broad-cause group 52 | Chalapati Rao, Alan D. Lopez, and Yusuf Hemed Figure 5.4 Summary of GBD Process for Estimating Cause- envelope of mortality, some of the disease-specific estimates Specific Mortality in African Countries described earlier were reduced to meet these envelope constraints. Cause-of-death WHO model model predictions life table of GBD group proportions by age RESULTS AND DISCUSSION Age-specific Distinct differences exist between countries belonging to the Age-specific Age-specific GBD cause group mortality risk mortality estimate two epidemiological subregions, essentially divided by mortality estimate the level of HIV/AIDS-related mortality within them. Overall, we estimate that about 10.8 million deaths (table 5.9) occurred in the year 2002 in the region, or just Disease-specific UN population epidemiological under 20 percent of global mortality. The age structure of estimate estimates mortality roughly divides into 46 percent of deaths occur- ring before the age of 15 years, another 36 percent between the ages of 15 and 59 years, and the remaining 18 percent Final cause-specific at age 60 and above. mortality estimate It is not surprising, therefore, that five of the six esti- by age mated leading causes of mortality in Sub-Saharan Africa are Source: Authors. Table 5.9 GBD Estimates of Leading Causes of Death, by Sex, 2000 Cause of death in Total deaths Cause of death in Total deaths Cause of death in Total deaths all persons (10,778,044) (%) males (5,557,783) (%) females (5,220,261) (%) 1 HIV/AIDS 20.4 1 HIV/AIDS 19.4 1 HIV/AIDS 21.6 2 Malaria 10.1 2 Lower respiratory infections 10.3 2 Malaria 11.0 3 Lower respiratory infections 9.8 3 Malaria 9.3 3 Lower respiratory infections 9.3 4 Diarrheal diseases 6.5 4 Diarrheal diseases 6.6 4 Diarrheal diseases 6.3 5 Perinatal conditions 5.1 5 Perinatal conditions 5.8 5 Perinatal conditions 4.5 6 Measles 4.1 6 Measles 4.0 6 Measles 4.2 7 Cerebrovascular disease 3.3 7 Tuberculosis 3.8 7 Cerebrovascular disease 4.1 8 Ischemic heart disease 3.1 8 Ischemic heart disease 3.1 8 Ischemic heart disease 3.1 9 Tuberculosis 2.8 9 Cerebrovascular disease 2.6 9 Tuberculosis 1.8 10 Road traffic accidents 1.8 10 Road traffic accidents 2.4 10 Pertussis 1.6 11 Pertussis 1.6 11 Violence 1.9 11 Road traffic accidents 1.2 12 Violence 1.2 12 Pertussis 1.5 12 Maternal hemorrhage 1.1 13 COPD 1.1 13 War 1.5 13 Nephritis and nephrosis 1.1 14 Tetanus 1.0 14 COPD 1.3 14 Tetanus 1.0 15 Nephritis and nephrosis 0.9 15 Tetanus 1.0 15 Diabetes mellitus 1.0 16 Malnutrition 0.9 16 Malnutrition 1.0 16 Malnutrition 0.8 17 War 0.8 17 Drownings 0.9 17 Maternal sepsis 0.8 18 Syphilis 0.8 18 Syphilis 0.9 18 COPD 0.8 19 Diabetes mellitus 0.7 19 Nephritis and nephrosis 0.8 19 Syphilis 0.8 20 Drownings 0.6 20 Prostate cancer 0.7 20 Hypertensive heart disease 0.8 21 All other specific causes 23.2 21 All other specific causes 21.4 21 All other specific causes 23.2 Source: Authors. Note: COPD chronic obstructive pulmonary disease. Causes of Death | 53 Table 5.10 GBD Estimates of Leading Causes of Death in AFR D and AFR E, 2000 Cause of death in AFR D persons (4,634,295) Total deaths (%) Cause of death in AFR E persons (6,143,749)a Total deaths (%) 1 Malaria 11.6 1 HIV/AIDS 28.4 2 Lower respiratory infections 11.3 2 Malaria 8.9 3 HIV/AIDS 9.8 3 Lower respiratory infections 8.6 4 Diarrheal diseases 7.4 4 Diarrheal diseases 5.8 5 Perinatal conditions 6.1 5 Perinatal conditions 4.4 6 Measles 5.6 6 Cerebrovascular disease 3.1 7 Cerebrovascular disease 3.7 7 Tuberculosis 3.0 8 Ischemic heart disease 3.5 8 Measles 2.9 9 Tuberculosis 2.6 9 Ischemic heart disease 2.8 10 Pertussis 2.2 10 Road traffic accidents 1.6 11 Road traffic accidents 2.1 11 Violence 1.3 12 Tetanus 1.4 12 War 1.2 13 Violence 1.2 13 Pertussis 1.2 14 COPD 1.1 14 COPD 1.1 15 Malnutrition 1.0 15 Nephritis and nephrosis 0.8 16 Nephritis and nephrosis 1.0 16 Malnutrition 0.8 17 Syphilis 0.9 17 Syphilis 0.8 18 Drownings 0.8 18 Tetanus 0.7 19 Diabetes mellitus 0.8 19 Diabetes mellitus 0.7 20 Hypertensive heart disease 0.6 20 Congenital anomalies 0.6 21 All other specific causes 25.3 21 All other specific causes 21.3 Source: Authors. Note: COPD chronic obstructive pulmonary disease. a. AFR E subregion includes countries with high HIV prevalence. those that cause deaths at childhood ages, namely, infectious The threefold difference in rank and percentage of deaths diseases and conditions originating in the perinatal peri- caused by HIV/AIDS between the two regions translates od (also see table 5.10). About 20 percent of the estimated into a nearly fourfold difference in absolute numbers of 2.2 million deaths due to HIV/AIDS are also predicted to deaths. There is reasonable similarity in the rank order and occur in childhood. Although the rank order of tuberculosis magnitude of the other leading causes, when deaths due to is almost similar in males (7) and females (9), the estimated all ages are combined. The importance of war as a cause of number of deaths in males (about 210,000) is more than death at all ages in AFR E stands out, being ranked as the double that predicted in females (95,000). Similarly, road 12th leading cause. traffic accidents, chronic obstructive pulmonary disease, At childhood ages, the rank structure and proportion of and war were estimated to cause more than double the deaths from individual leading causes is almost similar in the number of deaths in males than in females. In the entire two regions, and between males and females, except for the region, maternal conditions are estimated to cause 4.4 per- HIV/AIDS subregional difference mentioned earlier. Overall, cent of deaths in females. the subregional proportions of child mortality are nearly A subregional disaggregation of the magnitude of cause- equal; 48 percent of deaths occur in AFR D, and 52 percent specific mortality sheds more light on the epidemiological occur in AFR E. Sex differentials too are slight, with 52 per- variation possible in Sub-Saharan Africa. Two-thirds of the cent of deaths in males and 48 percent in females. total mortality occurs in countries within the AFR E region, Table 5.11 shows the leading causes of child deaths for the which contains just over half (54 percent) of the regional two subregions together. As expected, estimated leading caus- population. A comparison of the rank structure of leading es in Sub-Saharan Africa are infectious diseases, perinatal causes of death between the two epidemiological subre- conditions, and malnutrition. Two important implications of gions, AFR D and AFR E, is shown in table 5.10. these observations are that childhood mortality remains a 54 | Chalapati Rao, Alan D. Lopez, and Yusuf Hemed Table 5.11 GBD Estimates of Leading Causes of Death at Age 0 to 14 Years, 2000 Cause of death in males (2,594,761) Total deaths (%) Cause of death in females (2,368,129) Total deaths (%) 1 Malaria 18.1 1 Malaria 22.0 2 Lower respiratory infections 17.2 2 Lower respiratory infections 14.2 3 Diarrheal diseases 12.7 3 Diarrheal diseases 12.6 4 Perinatal conditions 12.4 4 Perinatal conditions 9.8 5 HIV/AIDS 8.5 5 Measles 9.1 6 Measles 8.4 6 HIV/AIDS 9.0 7 Pertussis 3.3 7 Pertussis 3.6 8 Road traffic accidents 2.0 8 Tetanus 1.8 9 Tetanus 1.7 9 Protein-energy malnutrition 1.5 10 Protein-energy malnutrition 1.6 10 Road traffic accidents 1.3 11 All other causes 14.2 11 All other causes 15.0 Source: Authors. Table 5.12 GBD Estimates of Leading Causes of Death at Age 15 to 59 Years, Subregional Comparison, 2000 Cause of death in Cause of death in AFR D persons (1,352,164) Total deaths (%) AFR E persons (2,536,058) Total deaths (%) 1 HIV/AIDS 25.8 1 HIV/AIDS 54.1 2 Tuberculosis 6.7 2 Tuberculosis 5.3 3 Lower respiratory infections 5.6 3 Lower respiratory infections 3.2 4 Violence 3.5 4 War 2.7 5 Road traffic accidents 3.5 5 Violence 2.6 6 Cerebrovascular disease 3.1 6 Road traffic accidents 1.9 7 Ischemic heart disease 2.6 7 Cerebrovascular disease 1.8 8 Malaria 2.3 8 Ischemic heart disease 1.5 9 Syphilis 2.2 9 Maternal hemorrhage 1.3 10 Maternal hemorrhage 1.9 10 Malaria 1.2 Source: Authors. major cause of burden in Sub-Saharan Africa and that there is excess mortality. Clearly, the number of deaths from each of no difference in the magnitude of the burden between differ- the leading causes is significantly higher in AFR E, despite a ent populations within the region. somewhat similar rank structure of causes apart from When comparing mortality during adulthood, however, HIV/AIDS and war. there are sizable differences, both between the two subre- As observed at childhood ages, the difference between gions and, within each subregion, between males and mortality in males and that in females is minimal (52 to 48). females. The mortality differences between the two subre- The leading causes of mortality in each sex in AFR E are gions, as seen from the leading causes for both sexes com- shown in table 5.13. bined, are shown in table 5.12. Apart from the high burden of mortality due to infec- The most striking feature is that the estimated number of tious diseases, it is evident that injury-related causes of deaths at these ages in AFR E is nearly double that estimated death among males, pregnancy-related causes among for AFR D. Readers will recall that the number of childhood females, and cardiovascular diseases in both sexes are major deaths in the two regions is nearly equal. From table 5.12 issues to be dealt with. The high mortality due to HIV/AIDS it is clear that in AFR E, the HIV/AIDS epidemic and in in both sexes has a major bearing on the occurrence of or- specific countries, armed conflicts are the causes of the phanhood within these countries. Causes of Death | 55 Table 5.13 GBD Estimates of Leading Causes of Death in AFR E at Age 15 to 59 Years: Comparison between Males and Females, 2000 Cause of death in males (1,308,048) Total deaths (%) Cause of death in females (1,228,010) Total deaths (%) 1 HIV/AIDS 50.3 1 HIV/AIDS 58.1 2 Tuberculosis 7.4 2 Lower respiratory infections 3.8 3 War 4.8 3 Tuberculosis 3.2 4 Violence 4.2 4 Maternal hemorrhage 2.6 5 Road traffic accidents 2.8 5 Cerebrovascular disease 2.0 6 Lower respiratory infections 2.7 6 Maternal sepsis 1.9 7 Ischemic heart disease 1.8 7 Abortion 1.4 8 Cerebrovascular disease 1.7 8 Hypertensive disorders of pregnancy 1.4 9 Malaria 1.1 9 Malaria 1.4 10 Syphilis 1.0 10 Ischemic heart disease 1.2 Source: Authors. Table 5.14 GBD Estimates of Leading Causes of Death at Age 60 Years and Older: Comparison between Males and Females, 2000 Cause of death in males (953,391) Total deaths (%) Cause of death in females (973,542) Total deaths (%) 1 Ischemic heart disease 13.4 1 Cerebrovascular disease 17.0 2 Cerebrovascular disease 10.8 2 Ischemic heart disease 13.4 3 Lower respiratory infections 6.5 3 Lower respiratory infections 5.6 4 COPD 6.1 4 Nephritis and nephrosis 3.8 5 Prostate cancer 4.0 5 Diabetes mellitus 3.7 6 Tuberculosis 3.1 6 COPD 3.5 7 HIV/AIDS 2.9 7 Hypertensive heart disease 3.2 8 Nephritis and nephrosis 2.9 8 Diarrheal diseases 2.6 9 Diarrheal diseases 2.3 9 HIV/AIDS 2.3 10 Diabetes mellitus 1.9 10 Cervix uteri cancer 2.3 Source: Authors. Note: COPD chronic destructive pulmonary disease. The subregional differences in mortality among the of lower respiratory infections, tuberculosis, diarrheal dis- elderly is similar to that observed in children: 47 percent in eases, and HIV/AIDS among the leading causes of death in AFR D and 53 percent in AFR E. The leading causes of death both sexes, even among the elderly, clearly defines the at these ages are similar for the two mortality strata. For importance of communicable disease control in African convenience, the discussion here will focus on the difference countries. in the cause-of-death structure between the two sexes, aggregated for both regions together or, in other words, for Sub-Saharan Africa. CONCLUSION As expected, in both sexes, cardiovascular diseases are estimated as the leading causes of mortality among the eld- Estimating cause-specific mortality in Sub-Saharan Africa erly (table 5.14). Among males, chronic obstructive pul- poses a major epidemiological challenge. The availability monary disease and prostate cancer are other leading causes, of data and information on causes of death has increased whereas in women, kidney disorders and diabetes mellitus within the region, and the results lead us to a few important are major causes of mortality. Cirrhosis of the liver is observations. First, childhood mortality is a major cause of estimated as the 11th leading cause in males; the correspon- the high premature mortality rates in Africa, accounting for ding rank for women is taken by breast cancer. The presence nearly half the total mortality in the region.As a corollary, the 56 | Chalapati Rao, Alan D. Lopez, and Yusuf Hemed observed leading causes of childhood mortality--malaria, ing the Sample Vital Registration and Verbal Autopsy diarrheal diseases, measles, lower respiratory tract infections, method of obtaining information on causes of death for and conditions originating in the perinatal period--require national mortality estimation purposes. This is the most immediate attention. viable interim solution to meet requirements of data for Second, the young adult population in countries within both health policy as well as for monitoring the impact of the AFR E mortality subregion is at significant risk of health programs and interventions. The GBD method of possible premature death from HIV/AIDS, armed conflict estimation, with the extensive use of models and extrapola- and other forms of violence, road traffic accidents, tubercu- tions from specific epidemiological studies, may not be use- losis, and, among women, causes related to pregnancy and ful for the monitoring function. childbearing. Another area of data collection that warrants attention at Third, nearly a fifth of the mortality in the region occurs this stage is the function of physician certification of cause among individuals age 60 years and older. The proportion of death and the implementation of guidelines, based on the of deaths at these ages is much less than that observed in International Classification of Diseases, for coding and clas- developed countries. For instance, in Australia, over 80 per- sification of causes of death. Capacity building in these areas cent of deaths occur above the age of 60 years (ABS 2003). will assist national health information systems in providing However, in Sub-Saharan Africa, the absolute number of useful information on causes of death, at least in urban deaths itself--1.92 million--merits attention to the health areas, where a significant number of deaths would occur in needs of the elderly. At these ages, cardiovascular disease, hospitals or where the deceased may have been attended by chronic obstructive pulmonary disease, cancers (prostate a physician in the time immediately preceding death. cancer in males, cervix and breast cancer in females), and, Undoubtedly, there is vast uncertainty about causes of notably, infectious diseases are the major causes of death in Africa, but enough is currently known to prepare mortality. preliminary estimates, adhering to a rigorous scientific It is generally accepted that statistics from complete vital framework for evaluation of data quality and ensuring sub- registration systems are the "gold standard" for national stantial prudence in interpreting the findings. What data are mortality statistics. Although cause-of-death information available suggest that further, massive responses to the HIV from vital records is subject to some biases on account of epidemic are needed and that major communicable diseases quality of cause-of-death attribution within the system, there and maternal health require scaled-up investments. is ample global evidence to justify a reliance on such data for Prevention of injuries, in part from war, would contribute national cause-specific mortality analysis and estimation. much to the improvement of health and survival of young It is heartening to note that countries within the region adult males. But what is most urgently needed is investment are making special efforts to improve vital registration in cost-effective methods to monitor mortality if we are not systems. The process, however, involves huge resources, and to be similarly ignorant about health conditions in Africa can be expected to take decades before data of reasonable 10 years hence. quality from national vital registration systems will be available for such estimation exercises. It has been observed in Kenya, South Africa, and Zimbabwe that although the NOTES coverage of registration of vital events by age and sex can 1. AFR D (high child and high adult mortality): Algeria, Angola, Benin, be improved rapidly by instituting certain administrative Burkina Faso, Cameroon, Cape Verde, Chad, Comoros, Equatorial Guinea, reforms, obtaining information on the cause of death Gabon, The Gambia, Ghana, Guinea-Bissau, Liberia, Madagascar, Mali, remains elusive, on account of deaths occurring at home in Mauritania, Mauritius, Niger, Nigeria, São Tomé and Principe, Senegal, Seychelles, Sierra Leone, Togo. AFR E (high child and very high adult mor- the absence of medical attention in remote areas. tality): Botswana, Burundi, Central African Republic, Côte d'Ivoire, An alternative system that has been tested and found Democratic Republic of Congo, Eritrea, Ethiopia, Kenya, Lesotho, Malawi, Mozambique, Namibia, the Republic of Congo, Rwanda, South Africa, effective has been sample registration, as has been employed Swaziland, Uganda, Tanzania, Zambia, Zimbabwe. in India and China. A representative sample of villages 2. The Coale-Demeny system is a system of model life tables that pre- or population clusters is routinely monitored for vital dicts age-specific mortality rates based on two parameters only, the level of (usually child) mortality in any age group and some idea of the relation- events, and upon death, a formal verbal autopsy procedure ship between infant and child mortality, which define the family (N, S, E, is employed to ascertain its probable cause. Initiatives such W). The Brass logit system is a relational model life table system that pre- dicts the set of age-specific death rates from knowledge of any two rates as the Health Metrics Network (established by the WHO) (usually child and adult ages 15­59) and the choice of a survival curve as are currently devising a framework for testing and establish- standard. Causes of Death | 57 REFERENCES Adequate Vital Registration Systems. International Journal of Epidemiology 32 (6): 1041­51. ABS (Australian Bureau of Statistics). 2003. Deaths Australia. Vol. 3302. Murray, C. J. L., B. D. Ferguson, A. D. Lopez, M. Guillot, J. A. Salomon, Annual report. Canberra: ABS. and O. Ahmad. 2003. "Modified Logit Life Table System: Principles, Adetunji, J. A., C. J. Murray, and T. Evans. 1996. "Causes of Death in Africa: Empirical Validation, and Application." Population Studies 57 (2): A Review." Paper presented at a meeting of the Population Association 165­82. of America, New Orleans, May. Murray, C. J. L., G. King, A. D. Lopez, N. Tomijima, and E. G. Krug. 2002. Black, R. E., S. S. Morris, and J. Bryce. 2003."Where and Why Are 10 Million "Armed Conflict As a Public Health Problem." British Medical Journal Children Dying Every Year?" Lancet 361: 2226­34. 324 (7333): 346­49. Brass, W. 1968. Demography of Tropical Africa. Princeton, NJ: Princeton Murray, C. J. L., and A. D. Lopez, eds. 1996. The Global Burden of Disease. University Press. Cambridge, MA.: Harvard University Press. Cliff, J., J. Sacarlal, O. Augusto, A. Nóvoa, M. Dgedge, G. Machatine, and Preston, S. H. 1976. Mortality Patterns in National Populations: With H. Cossa. 2003. Estudos das principais causas de morte registadas, nas Special Reference to Recorded Causes of Death. New York: Academic citades des Maputo, Beira, Chimoio e Nampula, em 2001. Maputo City, Press. Mozambique: Ministerio de Saude. Ravaoarimalala, C., R. Andriamahenina, B. Ravelojaona, D. Rabeson, Crowcroft, N. S., C. Stein, P. Duclos, and M. Birmingham. 2003. "How Best J. Andriamiadana, J. F. May, F. Behets, and A. Rasamindrakotroka. 1998. to Estimate the Global Burden of Pertussis?" Lancet Infectious Diseases "AIDS in Madagascar. II. Intervention Policy for Maintaining Low 3: 413­18. HIV Infection Prevalence." Bulletin de la Société de Pathologie Exotique. Deichmann, U. 1996. African Population Database. Digital Database and 91 (1): 71­73. Documentation. National Center for Geographical Information and Salomon, J. A., and C. J. L. Murray. 2002. "The Epidemiologic Transition Analysis, University of California, Santa Barbara. http://www.glowa- Revisited: Compositional Models for Causes of Death by Age and Sex." volta.de/cd_v3.1/landuse/populat/africa.htm. Population and Development Review 28 (2): 205­28. Dorrington, R., D. Bourne, D. Bradshaw, R. Laubscher, and I. Timaeus. Sankaranarayan, R., R. J. Black, and D. M. Parkin. 1998. Cancer Survival in 2001. The Impact of HIV/AIDS on Adult Mortality in South Africa. Cape Developing Countries. IARC Scientific Publications 145. Lyons: IARC Town: Medical Research Council. Press. Feachem, R. G., and D. T. Jamison, eds. 1991. Disease and Mortality in Sub- Shibuya, K., C. D. Mathers, C. Boschi-Pinto, A. D. Lopez, and C. J. Murray. Saharan Africa. New York: Oxford University Press. 2002. "Global and Regional Estimates of Cancer Incidence and Feachem, R. G., D. T. Jamison, and E. R. Bos. 1993. "Changing Patterns of Mortality by Site. II. Results of the Global Burden of Disease 2000." Disease and Mortality in Sub-Saharan Africa." In Disease and Mortality BMC Cancer 2 (1): 37. in Sub-Saharan Africa, ed. R. G. Feachem and D. T. Jamison. New York: New York University Press. Snow, R. W., M. Craig, U. Deichmann, and K. Marsh. 1999. "Estimating Mortality, Morbidity, and Disability Due to Malaria among Africa's Ferlay, J., F. Bray, P. Pisani, and D. M. Parkin. 2001. Globocan 2000: Cancer Non-Pregnant Population." Bulletin of the World Health Organization Incidence, Mortality and Prevalence Worldwide. Version 1. IARC 77 (8): 624­40. CancerBase 5. Lyons: IARC Press. Snyder, J. D., and M. H. Merson. 1982. "The Magnitude of the Global Hill, K., C. AbouZahr, and T. Wardlaw. 2001. "Estimates of Maternal Problem of Acute Diarrhoeal Disease: A Review of Active Surveillance Mortality for 1995." Bulletin of the World Health Organization 79 (3): Data." Bulletin of the World Health Organization 60: 604­13. 182­98. Stein, C. E., M. Birmingham, M. Kurian, P. Duclos, and P. Strebel. 2003. Kosek, M., C. Bern, and R. Guerrant. 2003. "The Global Burden of "The Global Burden of Measles in the Year 2000--A Model That Diarrhoeal Disease, as Estimated from Studies Published between 1992 Uses Country Specific Indicators." Journal of Infectious Diseases 187 and 2000." Bulletin of the World Health Organization 81 (3): 197­204. (Suppl. 1): S8­14. Kowal, P., C. Rao, and C. D. Mathers. 2003. Report on a WHO Workshop: Minimum Dataset on Ageing and Adult Mortality in Sub Saharan Africa. United Nations Population Division. 1998. World Population Prospects: Geneva: WHO. The 1998 Revision. New York: United Nations. Lopez, A. D., O. B. Ahmad, M. Guillot, B. D. Ferguson, J. A. Salomon, Waltisperger, D. M., P. Cantrelle, and O. Ralijaona. 1998. La mortalite C. J. L. Murray, and K. H. Hill. 2002. World Mortality in 2000: Life Tables a Antananarivo de 1984 a 1995. Paris: Centre Population et for 191 Countries. Geneva: WHO. Développement. Mathers, C. D., K. Shibuya, C. Boschi-Pinto, A. D. Lopez, and C. J. Murray. WHO (World Health Organization). 2003. World Health Report 2003-- 2002. "Global and Regional Estimates of Cancer Incidence and Shaping the Future. Geneva: WHO. Mortality by Site. I. Application of Regional Cancer Survival Model to WHO (World Health Organization) and UNAIDS (Joint United Nations Estimate Cancer Mortality Distribution by Site." BMC Cancer 2 (1): 36. Programme on HIV/AIDS). 2002. "Improved Methods and Mathers, C. D., C. Stein, D. Mafat, C. Rao, M. Inoue, N. Tomijima, Assumptions for Estimation of the HIV/AIDS Epidemic and Its C. Bernard, A. D. Lopez, and C. J. L. Murray. 2002. "Global Burden Impact: Recommendations of the UNAIDS Reference Group on of Disease 2000. Version 2: Methods and Results." GPE discussion Estimations, Modeling and Projections." AIDS 16: W1­W14. paper 50, WHO, Geneva. Williams, B. G., E. Gouws, C. Boschi-Pinto, J. Bryce, and C. Dye. 2002. Morris, S. S., R. E. Black, and L. Tomaskovic. 2003. "Predicting the "Estimates of World-Wide Distribution of Child Deaths from Acute Distribution of Under-Five Deaths by Cause in Countries without Respiratory Infections." Lancet 2: 25­32. 58 | Chalapati Rao, Alan D. Lopez, and Yusuf Hemed Chapter 6 Population and Mortality after AIDS Rodolfo A. Bulatao The acquired immune deficiency syndrome (AIDS) affects APPROACH AND DATA population size and composition in several ways. In partic- ular age groups, deaths increase directly from AIDS and may The projections of demographic impact to be considered also increase indirectly, as orphans, for instance, face higher come from three agencies: the Population Division of the mortality risks. Fertility can be affected, not only biologically United Nations (UN), the U.S. Census Bureau, and the but also from changes in sexual behavior. Communities may World Bank. Each agency has produced population projec- be weakened, and migration may alter the geographic distri- tions for most countries for some time. Up to 50 Sub- bution of the population. Saharan countries or territories are covered, although the This chapter draws on the work of agencies that pro- smallest ones are left out in some data series. For at least a duce global population projections to discuss the overall decade, these agencies have explicitly incorporated the effect population effect of AIDS in Sub-Saharan African coun- of AIDS in selected countries. In general, mortality due to tries. It does not attempt to elucidate all the mechanisms AIDS is added to a life table for a given country, and then involved--each of which deserves separate attention. overall mortality is projected into the future together with Instead, the chapter focuses on the broad demographic other vital parameters in order to project population. impact. Decades after the start of the epidemics, estimates Possible effects of AIDS on fertility (for example, Zaba and of this impact generally ignore all subtle and indirect effects Gregson 1998) and migration, except for indirect effects of and cover only the additional mortality directly from AIDS. mortality change, are not modeled. Even this impact is highly uncertain. There is general agree- The projections vary not only because of the mortality ment on substantial impact but no consensus on how sub- assumptions but also because of assumptions about initial stantial and long lasting it is. population size and composition and fertility and migration 59 trends. These differences are not detailed here. A recent Projections (UNAIDS Reference Group 2002), in World description and results are in United Nations 2003a, 2003b, Population Prospects: The 2004 Revision (United Nations 2004; for the other two sets of projections, recent descrip- 2004, 136­79). Calculations start with estimates of HIV tions are difficult to find, but one might consult earlier prevalence over time, mainly estimates for pregnant women descriptions in McDevitt 1999, Bos and colleagues 1994, and from antenatal clinics. The future trend of AIDS mortality National Research Council 2000. Results are in Stanecki is basically an extrapolation, with appropriate adjustments, 2004 and World Bank 2004. from this trend. HIV prevalence estimates as of 2001 are Although the approach to incorporating AIDS mortality taken from UNAIDS Reference Group 2002. For Djibouti, in the projections is similar across agencies, important details Gabon, Guinea, and Liberia, UNAIDS (2002) reports no are different. Agencies differ in how initial levels of mortality estimate, and the UN Population Division makes its own. are estimated and projected and how AIDS mortality is The source of estimates prior to 2001, used to establish a projected. trend, is not specified, but is presumably the database main- tained by the U.S. Census Bureau and similar sources. Although an attempt is made to match the UNAIDS (2002) United Nations Population Division estimate for 2001, fitting a trend sometimes leads to an HIV The Population Division of the UN seeks to apply country prevalence figure that varies from it, by a maximum of 2 per- life tables where available but mostly uses UN model life centage points (for Botswana) or 12 percent (for Uganda). tables. These are projected with reference to life expectancy, with gains expected to diminish as life expectancy rises. In U.S. Census Bureau the medium scenario (the only one considered here), the annual gain is 0.4 years for males and 0.5 years for females The U.S. Census Bureau follows a similar but not identical from a life expectancy level for both sexes combined of procedure (Stanecki 2004). Usually, a life expectancy target 60 years, but the gain shrinks to 0.1 years for males and is set in the future without considering AIDS, and logistic 0.2 years for females when life expectancy reaches 75 years. curves represent the trend to this target, with appropriate AIDS mortality is added to the life tables for 38 of the 50 life tables generated to match it. AIDS mortality is added countries or territories in Sub-Saharan Africa. The remain- for 37 Sub-Saharan countries. In contrast to the countries der, with reported low prevalence of the human immunod- covered by the UN Population Division, the Census Bureau eficiency virus (HIV) or no data from UNAIDS, are island does not introduce AIDS-specific deaths for Equatorial countries, except for Mauritania, Niger, Senegal, and Guinea, The Gambia, and Sudan but does cover Niger and Somalia. A criterion of HIV prevalence no higher than Senegal. The criterion for inclusion is national HIV preva- 1 percent appears to be used in excluding countries, but the lence above 1 percent (as estimated by UNAIDS for 1999), size of the affected population is also taken into account. HIV prevalence of 5 percent or more in low-risk urban pop- (This is noticeable mainly outside Sub-Saharan Africa, ulations, or a prevalence trend that suggests the latter level because AIDS mortality is also added for China.) will soon be reached. To add AIDS mortality, a separate projection of Instead of modeling the AIDS epidemic in each country HIV/AIDS in the population is necessary. The procedure separately, the bureau takes a shortcut. It defines five epi- begins with reported levels of HIV prevalence among adults. demic scenarios, progressively more severe, ranging from a With appropriate assumptions, the UN back-calculates the "low" to a "super high" epidemic, each starting at an indefi- incidence of HIV and distributes cases by sex and age. From nite date. These scenarios are developed using a program the size of the infected population and the remaining, labeled iwgAIDS (Stanley et al. 1991), which models the susceptible population, the UN estimates the trend in sub- spread of HIV infections, the development of AIDS, and sequent infections. It derives mortality associated with AIDS subsequent deaths in a population. Each scenario incorpo- given assumptions about the progression of the infection rates the effect of increased condom use and shows HIV and adds additional mortality from perinatal infections. Life prevalence eventually plateauing and declining. tables are adjusted for the additional mortality. Each country is assigned to a scenario, or more precisely The Population Division has described the procedure for to an interpolated scenario between an assigned pair of estimating AIDS mortality, which follows the recommenda- these five. The assignment is made by matching the reported tions of the Joint United Nations Programme on HIV/AIDS trend in HIV prevalence in urban areas to scenario trends. (UNAIDS) Reference Group on Estimates, Modelling and Total country prevalence is then used to determine dates for 60 | Rodolfo A. Bulatao the epidemic. It is assumed that HIV prevalence will peak in data from Tanzania, South Africa, and Zimbabwe (Lopez 2010 and AIDS mortality will decline to zero by 2070. AIDS et al. 2001), is imposed, and AIDS mortality is then added deaths are therefore added to the life table, from the projected to the life table. epidemic, up to 2070, at which time the life table reverts to In projecting mortality forward from the initial life tables what it would be without the epidemic. taken from the WHO, the World Bank does not attempt to Uganda is treated as a special case, because of a reported model the HIV/AIDS epidemics in any way, but simply decline in prevalence from the mid-1990s, which violates assumes that their effect will gradually be reduced to zero by the assumption of a 2010 peak. Separate epidemics are mod- 2020 at the latest (depending on the country), at which time eled: a high one up to 1995 and a low one from 2005 on, country mortality trends resume a standard pattern used in with an interpolated epidemic in the intervening period. the Bank's projections. The initial HIV national prevalence levels used are simi- lar to those adopted by UNAIDS, relying heavily on report- Comparisons ed infections among pregnant women. In this, the bureau projections resemble those of the UN Population Division. Counterfactual projections that include no effect of AIDS However, the future trend imposed on prevalence is unusual. are useful in assessing impact. Only the UN Population The year 2010 represents a later peak to the epidemics than Division provides detailed results for such a scenario. The projected by the UN, in most cases. Peak incidence, in bureau has run such scenarios and provides brief descrip- 55 percent of the Sub-Saharan cases in the UN projections, tions but quite limited numerical information on the results is before 1995, and in 95 percent of the cases before 2000. (Stanecki 2004). The World Bank has no such scenario. Although prevalence tends to peak later than incidence, this Most comparisons to be made, therefore, will rely on the UN is unlikely to adequately account for the long delays between no-AIDS scenario. UN incidence peaks and Census Bureau prevalence peaks. It is not obvious from the approach taken in each projec- tion set which should show the greatest current impact of AIDS on population and mortality. For future impact, how- World Bank ever, a possible order emerges. Because the Census Bureau World Bank mortality projections begin with life tables that assumes a relatively late peak to the epidemics, one might already incorporate AIDS. However, these life tables, devel- expect greater demographic impact in its projections. oped by the World Health Organization (WHO) for 2000, Because the World Bank assumes a fairly direct decline in are constructed in a familiar way, with AIDS mortality being mortality impact, with the elimination of this impact by added to mortality from all other causes. 2020 instead of the Census Bureau's projection of 2070, one The WHO takes estimates of child mortality (5q0) and might expect these projections to show the least future adult mortality (45q15) and expands them into life tables for impact. This depends crucially on specific parameters, how- each country, using the Brass logit approach (Murray et al. ever, and the results of comparisons need not be uniform 2000; Lopez et al. 2001). The child mortality estimates were across countries. selected from survey and census estimates after systematic review of available statistics (Ahmad, Lopez, and Inoue 2000). The adult mortality estimates were more problematic. EFFECTS ON POPULATION SIZE, DECOMPOSED For Africa, plausible national estimates in the literature prior to 1990 (to exclude the effect of AIDS) were used, supple- Population trends taking AIDS into account are illustrated mented where necessary with UN estimates (United Nations for 2000­50 in figure 6.1, which compares results from the 1998), and projected forward to 1999 and, later, 2000. three agencies. An additional line represents the special UN To add AIDS mortality, the WHO begins with reported scenario with no AIDS mortality. For Sub-Saharan Africa as HIV prevalence (based, as usual, on pregnant women) and a whole, the three projections of total population are well works backward to HIV incidence and then forward to below the no-AIDS scenario practically from 2000 and fairly obtain an overall AIDS mortality level. (The researchers close to each other up to about 2015. By 2020, some caution, based on preliminary work for Zimbabwe, that divergence appears, with the Census Bureau projection "these models may overestimate the level of the epidemic" 0.1 percent above the UN projection and the World Bank [Salomon, Gakidou, and Murray 1999, 8], a point we projection 2.5 percent below it. Divergence increases as the return to below.) An age pattern of deaths, based on limited projections lengthen, so that by 2050 the Census Bureau Population and Mortality after AIDS | 61 Figure 6.1 Projected Population, Sub-Saharan Africa and Three Selected Countries (thousands) Sub-Saharan Africa Nigeria 2,000,000 400,000 1,800,000 350,000 1,600,000 300,000 1,400,000 population population 250,000 1,200,000 200,000 projected 1,000,000 projected 150,000 800,000 600,000 100,000 2000 2010 2020 2030 2040 2050 2000 2010 2020 2030 2040 2050 year year Benin Swaziland 18,000 2,500 2,250 16,000 2,000 14,000 1,750 population 12,000 population 1,500 1,250 10,000 projected projected 1,000 8,000 750 6,000 500 2000 2010 2020 2030 2040 2050 2000 2010 2020 2030 2040 2050 year year UN (no AIDS) UN U.S. Census World Bank Sources: Stanecki 2004; United Nations 2004; World Bank 2004. projection is 8.6 percent above the UN projection and the affect population size, its effects are entangled with, and may World Bank projection is 9.4 percent below. This is the be overwhelmed by, assumed change (or lack of change) in reverse of the relative impact one would expect from com- other parameters, particularly fertility. paring methodologies. Fertility assumptions are indeed quite variable across Three countries are shown in figure 6.1: Nigeria, with the the projection sets. The UN projection and the no-AIDS largest regional population and an HIV prevalence rate scenario, but not the other two projections, have similar intermediate for the region (5.4 percent infected of adults 15 total fertility rates by design. World Bank estimates of total to 49 years old, as reported by UNAIDS [2004] for 2003); fertility in Sub-Saharan countries for 2000­05 range from Benin, with low HIV prevalence, for the region, of 1.9 per- 21 percent below to 23 percent above UN estimates, with a cent in 2003; and Swaziland, with one of the highest HIV tendency to be lower rather than higher. Census Bureau esti- prevalence rates, for the region or the world, of 38.8 percent mates, in contrast, range from 37 percent lower to 53 percent in 2003. For these individual countries, greater divergence in higher than UN estimates, with a decided tilt toward being projections is evident than for the region as a whole, some higher. The variation tends to increase, in percentage terms, of it clearly not due to AIDS mortality. The Census Bureau, for later years, at least up to around 2025. Higher bureau and for instance, projects a larger population than in the no-AIDS lower Bank fertility projections appear to counteract the projection for Nigeria by 2035, as well as a larger population effect on projected population of longer-lasting epidemics for Benin all the way back to 1990. Although AIDS does for the Census Bureau than for the Bank. 62 | Rodolfo A. Bulatao Table 6.1 Effects of Demographic Factors in Producing Differences between Projections 2020 population 2050 population Comparison Base Net Base Net and indicator Total population Births Deaths migrants Total population Births Deaths migrants UN vs. no AIDS Sub-Saharan Africa 0.890 0.976 1.014 0.901 0.998 0.830 0.976 1.026 0.923 0.989 Country maximum 1.000 1.000 1.052 1.000 1.004 1.000 1.000 1.105 1.000 1.000 Country minimum 0.591 0.907 1.000 0.624 0.981 0.370 0.907 1.000 0.646 0.951 U.S. Census vs. no AIDS Sub-Saharan Africa 0.891 0.971 1.020 0.910 0.988 0.901 0.971 1.067 0.930 0.978 Country maximum 1.252 1.161 1.232 1.043 1.091 1.484 1.161 1.242 1.027 1.098 Country minimum 0.516 0.637 0.758 0.633 0.803 0.322 0.637 0.844 0.620 0.783 World Bank vs. no AIDS Sub-Saharan Africa 0.868 0.985 0.983 0.898 0.998 0.752 0.985 0.984 0.922 0.989 Country maximum 1.034 1.122 1.115 1.029 1.070 1.100 1.122 1.090 1.013 1.034 Country minimum 0.660 0.791 0.861 0.709 0.916 0.515 0.791 0.850 0.798 0.955 Source: Calculated from data in United Nations 2004; Stanecki 2004; and World Bank 2004. See note 1 at the end of the text. Note: An effect of 0.89, for instance, means that the specified factor by itself reduces one projected population to 89 percent of the comparison population. Maximum and minimum effects are across 48 countries (the Seychelles and Saint Helena having incomplete data). Ideally the effect of fertility assumptions would be dis- small for the region as a whole, at 1.4 percent in 2020, rising counted by running parallel scenarios with and without to 2.6 percent by 2050. It is, however, positive for each coun- AIDS, as the UN has done. Such scenarios are not available, try where some effect of AIDS is modeled. Since the same at least not in sufficient detail, for the other two sets of total fertility is assumed country by country in the UN projections, but we do attempt to separate the effects of the projection and its no-AIDS scenario, this result deserves fertility and mortality assumptions. We decompose the ratio some explanation. of populations from two separate projections into multi- Crude birth rates differ between the projections, as fig- plicative factors representing differences in the assumed ure 6.2 illustrates. For the region as a whole, the difference base population and in fertility, mortality, and migration is relatively slight, but for countries severely affected by assumptions.1 This decomposition does not specifically AIDS, the difference can be much larger, as in the projec- identify AIDS impact, but it at least helps us separate out the tion for Botswana. AIDS raises the crude birth rate because fertility and mortality effects. of its effect on the age structure. Figure 6.3 shows the dis- Table 6.1 shows the results for 2020 and 2050. The com- tribution by age group of women of reproductive age in parison between the UN projection and the no-AIDS sce- Botswana. Whereas, without AIDS, the age distribution is nario (also from the UN) has some interesting points. The projected, over 50 years, to gradually become rectangular projected total population for 2020 is 89 percent of the over the range of 15 to 49 years, with AIDS it remains projection without AIDS (as the first "Total" column skewed toward those under 35. This translates into a higher shows). Therefore, the overall reduction in population due crude birth rate because fertility is higher among women to AIDS is 11 percent by 2020, 17 percent by 2050. Most of under 35 years. The fertility advantage of younger women is this is due to mortality, but other effects also contribute. Part assumed to decline over time, but the decline does not is due to differences in the base population, the population entirely offset the relative increase, when AIDS is taken into assumed for the start of the projection in 2000. The UN esti- account, in proportions of younger women. The effect is mates that AIDS had already produced some demographic entirely distributional and does not involve any assump- effect before 2000, reducing regional population by over tions about the biological impact of HIV, which, if it were 2 percent. The largest such effect across countries is 9 per- taken into account, might moderate the result somewhat. cent, for Zimbabwe. AIDS does not increase the number of births, but these Births also have an effect--a positive one in the UN results indicate that it does increase the ratio of births to the projection, in contrast to the no-AIDS scenario. The effect is population. Population and Mortality after AIDS | 63 Figure 6.2 Projected Crude Birth Rate, Sub-Saharan Africa Figure 6.3 Distribution of Women of Reproductive Age in the and Botswana UN Projection and the No-AIDS Scenario, Botswana Sub-Saharan Africa 25 45 20 40 rate 15 35 birth women 30 % 10 crude 25 5 projected 20 0 15­19 20­24 25­29 30­34 35­39 40­44 45­49 15 2000 2010 2020 2030 2040 2050 age year no AIDS UN Botswana 2000 2025 2050 2000 2025 2050 35 Source: United Nations 2004. 30 rate birth 25 crude 20 Figure 6.4 Mortality Effects on Population, Relative to the projected 15 No-AIDS Scenario, in Different Projections 10 2020 2000 2010 2020 2030 2040 2050 110 year UN (no AIDS) UN U.S. Census World Bank 100 Sources: Stanecki 2004; United Nations 2004; World Bank 2004. effect 90 80 mortality The Census Bureau also shows a generally positive births % 70 effect. Part of this, as for the UN, may be due to the effect of AIDS on the age structure, but a larger part is simply higher assumed fertility than in the no-AIDS scenario. As expected, 60 65 70 75 80 85 90 95 100 % UN mortality effect the births effect is negative for the World Bank. The direct effect of mortality on population size in the 2050 projections is of course negative and much larger in the 110 absolute than the effect of fertility, implying a reduction in 100 population size for the region, if the UN projection is com- pared with the no-AIDS scenario, of 10 percent by 2020 and effect 90 8 percent by 2050. For the region as a whole, the mortality effect is roughly comparable, if the U.S. Census Bureau 80 mortality % projections or the World Bank projections are compared with 70 the same standard. We expected the Census Bureau to show a stronger mortality effect than the World Bank, given the 60 65 70 75 80 85 90 95 100 assumption of longer-lasting epidemics, but the reverse is the % UN mortality effect case, except when the most affected countries are considered. U.S. Census World Bank UN Figure 6.4 shows the mortality effects for individual countries--with the most affected countries toward the Sources: Calculated from Stanecki 2004, United Nations 2004, and World Bank 2004. 64 | Rodolfo A. Bulatao lower left--when assessed for 2020 and for 2050. Although LIFE EXPECTANCY the agencies generally agree about which countries are more affected by AIDS mortality, their estimates of the degree of Looking at other mortality indexes does not help further this effect vary, particularly for the most affected countries. specify the AIDS effect but does provide additional perspec- For countries where the UN estimates that AIDS mortality tives on the severity of the impact. Figure 6.5 shows some reduces population in 2020 by at least 20 percent, the projected life expectancy trends. For Sub-Saharan Africa as Census Bureau estimates that the population reduction on a whole, the UN Population Division indicates that by average is 5 percentage points less, and the World Bank 2000­05 life expectancy had already fallen 8.8 years short estimates that it is 8 percentage points less. For countries of what it would be without AIDS and that the relative where the UN estimates a population reduction of at least deficit will grow to a maximum of 10.4 years by 2010­15 20 percent by 2050, the Census Bureau estimates an average before beginning to shrink slowly. The Census Bureau reduction that is 3 percentage points less, and the World suggests that the current deficit is slightly smaller, and the Bank a reduction of 13 percentage points less. World Bank suggests it will shrink slightly faster. For the These comparisons involve not only AIDS mortality but region as a whole, there is not great disagreement about life mortality generally. It is likely, however, that the differences expectancy trends. have to do mainly with variation in AIDS mortality than For individual countries, however, the differences among in mortality from all other causes, although this cannot be projections can be substantial. The UN projections start (in established. At any rate, the results fail to substantiate any 2000) with a much higher estimate of life expectancy for general agreement about the demographic effect of AIDS Nigeria than the other two projections, a gap that is generally mortality. maintained over time. For Benin, the Census Bureau Figure 6.5 Projected Life Expectancy, Sub-Saharan Africa and Three Selected Countries Sub-Saharan Africa Nigeria 75 75 70 70 65 65 60 60 55 55 expectancy 50 expectancy 50 life 45 life 45 40 40 35 35 30 30 2000 2010 2020 2030 2040 2000 2010 2020 2030 2040 year year Benin Swaziland 70 80 70 65 60 60 50 expectancy expectancy 55 life life 40 50 30 45 20 2000 2010 2020 2030 2040 2000 2010 2020 2030 2040 year year UN (no AIDS) UN U.S. Census World Bank Sources: Stanecki 2004; United Nations 2004; World Bank 2004. Population and Mortality after AIDS | 65 Figure 6.6 HIV Prevalence and Maximum Loss in Life death rate tends to be a few years earlier than the greatest Expectancy in Alternative Projections loss in life expectancy. 5 The decline in life expectancy affects women more than men. Worldwide, in the period 2000­05, women had a 0 4.3 year advantage in life expectancy. In Sub-Saharan Africa, 5 their advantage would have been 3.1 years without AIDS expectancy 10 life and would have stayed essentially constant up to 2050 (in in 15 the no-AIDS scenario). Because of AIDS, their advantage lost 20 has fallen to 1.9 years, according to the UN projection, and years 25 will fall further to 0.6 years for most of the period 2010­25, 30 before recovering slowly, to barely more than 1.5 years by maximum 35 2050. The Census Bureau shows less of a change, with the 40 female advantage at 2.2 years in 2000­05, falling under 0 5 10 15 20 25 30 35 40 HIV adult prevalence (UNAIDS) two years during 2005­15, and recovering more quickly to about four years by 2050. (The World Bank does not provide UN U.S. Census World Bank these data.) Sources: Calculated from Stanecki 2004, United Nations 2004, and World Bank 2004. Country by country, projected trends in the female advantage can look strikingly different (figure 6.7). projects a substantial fall in life expectancy, unlike the other Although there are some commonalities in country pat- two agencies. For Swaziland, the decline in projected life terns, the UN and the Census Bureau appear to disagree expectancy up to about 2010­15 is greatest for the UN, about many aspects of these trends: what the female advan- which has it falling to 30 years. The decline is smallest for tage is initially, when and how steeply it declines, and if and the World Bank, for which the minimum, in 2005­10, is when it begins to increase again. The Census Bureau gener- 43 years. Subsequent trends for Swaziland are also quite ally assumes an earlier decline in the female advantage than different: the UN projects a long, slow recovery; the World the UN but a somewhat shallower decline at lower HIV Bank sees a quick, rapid recovery; and the Census Bureau prevalence levels. In both sets of projections, the female finds a delayed but accelerating recovery. advantage falls by a year for every 4 to 5 percentage point How far life expectancy falls depends on current esti- increase in initial HIV prevalence. This can turn the female mates of HIV prevalence. Figure 6.6 shows the maximum advantage into a disadvantage. fall in life expectancy by country, when each set of projec- Where HIV prevalence reaches 20 percent or more, tions is contrasted with the no-AIDS scenario. In the UN female life expectancy usually falls, at some point in the pro- projection, a country loses four years of life expectancy jection, at least two years below male life expectancy. Most when HIV prevalence is 2.5 percent among adults, and a little of these countries are in southern Africa, or in the slightly more than one additional year of life expectancy for every broader Southern African Development Community 1 percentage point rise in HIV prevalence.2 The U.S. Census (SADC). The UN always shows some recovery by 50 years, Bureau projects a similar result (when the comparison is although in the worst case, Botswana, the female disadvan- made to the UN no-AIDS scenario), and the World Bank tage, at 4.5 years, is still extreme at the end of this period. projects a somewhat less serious effect, particularly at high The Census Bureau, in contrast, shows either a quick and HIV levels. The greatest effect on life expectancy is most substantial recovery, as is shown in figure 6.7 for Zimbabwe, often seen in the period 2015­20. For the UN, the period or an apparently unarrested fall, as for Namibia, where the may be five years earlier or later, but for the Census Bureau, female disadvantage exceeds seven years by 2050. Long-term there is less temporal variation. For the World Bank, however, trends, particularly where HIV prevalence is high, are a par- the period of maximum deficit is 2005­10, and this is fairly ticular area of disagreement. constant across countries. The changes in life expectancy, and in the female advan- The crude death rate is affected similarly. The maximum tage, are the result, of course, of rising adult mortality. The increase in the rate is about 7 per 1,000 people when HIV estimates of adult mortality (45q15) used in the projections prevalence is 2.5 percent, rising to 25 per 1,000 when HIV are available only for the UN. These show the substantial prevalence is 35 percent. The greatest increase in the crude rise in mortality over projected trends without AIDS, as well 66 | Rodolfo A. Bulatao Figure 6.7 Female Advantage in Life Expectancy, Sub-Saharan Africa and Three Selected Countries Sub-Saharan Africa Benin 4 5 4 3 (years) (years) 3 2 advantage advantage 2 female 1 female 1 0 0 1990 2000 2010 2020 2030 2040 2050 1990 2000 2010 2020 2030 2040 2050 year year Zimbabwe Namibia 6 6 4 4 2 2 (years) (years) 0 0 advantage advantage 2 2 4 female female 4 6 6 8 1990 2000 2010 2020 2030 2040 2050 1990 2000 2010 2020 2030 2040 2050 year year no AIDS UN U.S. Census Sources: Calculated from Stanecki 2004 and United Nations 2004. as the way female adult mortality catches up with and, in the sex ratio may rise 10 to 20 percentage points in the UN severe cases, passes male mortality (figure 6.8). The increases projection and slightly more in the Census Bureau projec- in the risk of death between ages 15 and 60, relative to the tion. As a result, the highest sex ratio in these projections situation without an AIDS epidemic, can be substantial. If would be about 120 males per 100 females. This is not HIV prevalence in 2001 is 10 percent, the maximum unprecedented. In 2000, five countries--all in the Persian increase in risk over the following two decades is, on average Gulf and none particularly affected by AIDS--had higher from a quadratic regression, 29 percentage points for males ratios, up to 190 males per 100 females. National sex ratios and 34 percentage points for females. If HIV prevalence is approaching 120 could be destabilizing, depending on the 30 percent, the increase in risk is 65 percentage points for flexibility, or lack of it, of particular cultures. males, 75 percentage points for females. At specific ages, sex ratios may become more extreme. Figure 6.9 illustrates the situation for Zimbabwe, one of the countries most affected by AIDS mortality. Over decades, Population by Sex and Age the imbalance between males and females resulting from The shrinking (and occasionally disappearing) female higher female AIDS mortality produces a hump in the sex advantage in life expectancy is projected to alter the sex ratio, which, in this case, reaches about 230 in the year 2030 ratio, although for national populations the effect is not for those age 50 to 54 years. This hump, as it rises, gradually large. For a handful of southern African or SADC countries, moves toward older and older ages and eventually begins to Population and Mortality after AIDS | 67 Figure 6.8 Adult Mortality (45q15) by Sex in UN Projections, Figure 6.9 Sex Ratios by Age from UN Projections, Zimbabwe Sub-Saharan Africa and Lesotho 250 Sub-Saharan Africa 1.0 200 0.8 150 ratio sex 0.6 100 q 15 45 0.4 50 0.2 0 0­4 5­910­14 15­19 20­24 25­29 30­34 35­39 40­44 45­49 50­54 55­59 60­64 65­69 70­74 75­79 80­84 85­89 90­94 95­99 100 0 age 1990 2000 2010 2020 2030 2040 2050 year 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 Lesotho 1.0 Source: United Nations 2004. 0.8 demographic transition--toward lower dependency, a stalled fertility decline is at least as effective an impediment. 0.6 q 15 45 0.4 Initial Assumptions 0.2 Is it possible to sort out the differences in projections and come to some reasonable conclusions about the likely 0 1990 2000 2010 2020 2030 2040 2050 demographic impact of AIDS? There is no way to determine year in advance which forecasts will be most accurate in males males (no AIDS) females females (no AIDS) the decades to come. The only useful approach to assessing the relative adequacy of the different projections, therefore, Source: United Nations 2004. would be to look at their methodology, especially the way they model the effects of HIV/AIDS. These models, however, subside. Having almost two-and-a-half males for every are quite complex, not completely transparent, and beyond female at older ages is unusual, but whether it is a serious the scope of this chapter to critique in detail. We therefore issue depends on the culture. settle for examining some of the initial assumptions that go Where age patterns are involved, the dependency ratio into these projections, particularly assumptions about adult might be a particular concern. A low dependency ratio, mortality and HIV prevalence. These two sets of parameters implying a relatively large workforce, could be an economic undergird projections of AIDS impact. advantage, but AIDS impedes progress toward a lower ratio For the World Bank projections, calculations of adult by increasing adult mortality more than infant and child mortality can be made from the WHO life tables for 2000 mortality. How great the effect is, though, is difficult to tell. (Lopez et al. 2002). Similar calculations can be made from In the countries most severely affected by AIDS, the depend- UN survivorship ratios; for 2000, we average estimates for ency ratio might rise from 50 or 60 dependents per 100 adults the periods 1995­2000 and 2000­05. (Similar data are not of working age to 70 or 80 dependents. Comparison of the available from the Census Bureau.) UN figures are plotted UN projection with the no-AIDS scenario suggests such a against WHO figures by country in figure 6.11. rise for various countries in southern Africa. However, the The UN and WHO adult mortality estimates generally exact level of the dependency ratio will be heavily influenced agree. On average, UN estimates are only 1 percent lower by future fertility, about which there appears to be little con- than WHO estimates. (The differences may be slightly sensus (figure 6.10). The apparent effect of AIDS is easily greater at high mortality levels.) Figure 6.11 also shows, swallowed up by differences in assumed fertility trends. however, another set of estimates of adult mortality, derived Although AIDS does delay the movement--typical in the by Timaeus and Jasseh (2004) from sibling histories and 68 | Rodolfo A. Bulatao Figure 6.10 Dependency Ratio, Sub-Saharan Africa and Three Selected Countries Sub-Saharan Africa Nigeria 100 100 90 90 80 80 ratio ratio 70 70 dependency 60 dependency 60 50 50 40 40 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 year year Benin Swaziland 100 100 90 90 80 80 ratio ratio 70 70 dependency 60 dependency 60 50 50 40 40 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 year year no AIDS UN U.S. Census World Bank Sources: Stanecki 2004; United Nations 2004; World Bank 2004. incorporating recent Demographic and Health Surveys The Timaeus and Jasseh (2004) estimates cannot be con- (DHS) data. These estimates, which cover seven countries, sidered an entirely reliable standard because of the limited appear consistently lower than those from the WHO and the data on which they are based and the assumptions necessary UN. (The only exception is for Zimbabwe, where the WHO in their calculation, such as the assumption of a common estimate, but not the UN estimate, is slightly lower.) The gap age pattern of mortality increase across countries (see also appears to be substantial. The probability of dying between chapter 4 in this volume). Nevertheless, the consistently ages 15 and 60, averaged without weighting across the seven lower levels of adult mortality they determine may suggest countries, is between 51 and 57 percent in the WHO and UN some overstatement in the UN and WHO mortality estimates (depending on which estimates and which sex is estimates, which could be tied to an overestimation in the involved). The averages for Timaeus and Jasseh's estimates, in modeling of the mortality impact of AIDS. contrast, are 44 percent for males and 34 percent for females. Without getting into the intricacies of these models, we Somewhat more countries can be compared for 1995, for can look at the estimates of HIV prevalence with which the which year Timaeus and Jasseh (2004) cover 20 countries. modeling starts. These are generally UNAIDS estimates for These estimates are mostly well below the UN estimates for 2001. The UN estimates, as noted earlier, begin with these 1995 (estimated as the average for 1990­95 and 1995­2000). estimates but may be adjusted in order to fit the prevalence The average probability of dying between ages 15 and 60 trend. The Census Bureau maintains the database that across the 20 countries is 36 percent for males and 29 per- UNAIDS uses in making its estimates and reports similar cent for females, according to Timaeus and Jasseh (2004). figures, at least for countries for which it has estimates. The According to the UN, the respective averages are 50 percent source of the WHO estimates is not specified but is presum- for males and 42 percent for females. ably the same. Population and Mortality after AIDS | 69 Figure 6.11 Consistency among Alternative Estimates of Figure 6.12 HIV Prevalence in 2001 and 2003 Current Adult Mortality (45q15) 40 Swaziland Males 35 Botswana 1.0 30 2003 Lesotho 0.8 25 South Africa Zimbabwe Namibia 20 0.6 Zambia prevalence 15 q 15 45 HIV 10 0.4 Cameroon Kenya 5 0.2 0 5 10 15 20 25 30 35 40 HIV prevalence 2001 0 0.2 0.4 0.6 0.8 Source: UNAIDS 2002, 2004. q from WHO life tables 45 15 Females points lower. Among the increases, only one country 1.0 (Swaziland) has a 2003 rate at least 1.5 points higher than in 2001. Between 1999 and 2001, in contrast, increases and 0.8 decreases in prevalence were more balanced, with increases 0.6 being slightly more likely and somewhat larger. q If the epidemics are indeed at a plateau or past their peak, 15 45 0.4 that may be uncomfortable for the agency projections. The Census Bureau projections are inconsistent with this idea, 0.2 since they assume that peak prevalence is not reached until 2010. Whether the UN projections are consistent with 0 0.2 0.4 0.6 0.8 plateauing epidemics is not clear. The UN reports that peak q from WHO life tables 45 15 incidence for the epidemics--not peak prevalence--was WHO UN Timaeus and Jasseh reached on average about 1994, which could be consistent with some decline in prevalence by 2001. Country by coun- Sources: Timaeus and Jasseh 2004; United Nations 2004; Lopez et al. 2002. try, however, there is no relationship between an earlier A problem with each of these projections is the 2001 assumed incidence peak for the UN and the amount of UNAIDS estimates themselves. In 2004, UNAIDS (2004) apparent prevalence decline in the UNAIDS estimates. Only revised these estimates, changing almost twice as many the World Bank projections, which effectively assume that downward as upward. The changes were sometimes large. AIDS-related mortality is already in decline, would be con- UNAIDS (2002) had previously reported "low" and "high" sistent with the interpretation of epidemics in at least early estimates to bracket their main estimates for 2001.In 10 cases, decline in the region. the new estimates were lower than the previous "low" esti- The correction of 2001 HIV prevalence rates is also mates. In only two cases were they higher than the previous uncomfortable for projections, which relied on the earlier "high" estimates. The adjustments may have been made to rates. The correction seems to have some basis. Table 6.2 ensure consistency with new estimates for 2003. These were compares UNAIDS estimates with estimates from DHS data generally lower than the earlier 2001 estimates, which might for five countries. Based mainly on antenatal clinic reports, have suggested epidemics on the decline. Instead, by lowering the earlier UNAIDS estimates were typically higher than the 2001 estimates, UNAIDS can speak of "stabilization" in HIV DHS estimates. Among these countries, rather high rates for prevalence in the region, which it attributes to a balance Kenya and Zambia were brought into the range of the DHS between deaths and new infections. estimates, and rates for the other three countries were also Figure 6.12 compares UNAIDS estimates of HIV preva- lowered. lence in 2003 with the earlier estimates for 2001 (UNAIDS The survey estimates are meant to cover the adult popu- 2002, 2004). Among the reductions, 15 countries have 2003 lation more comprehensively than sentinel surveillance rates that are at least 1.5 percentage points lower than the systems based on special populations, especially pregnant earlier 2001 estimates, three of them at least 5 percentage women. Survey estimates do have their own problems, such 70 | Rodolfo A. Bulatao Table 6.2 Estimates of HIV Prevalence among Adults Age 15 to 49, 2001­03 (percent) UNAIDS Demographic and Health Surveys Country, by prevalence 2001 2001 (rev.) 2003 Year Female Male Zambia 21.5 16.7 16.5 2001­02 17.8 12.9 Kenya 15.0 8.0 6.7 2003 8.7 4.6 Burkina Faso 6.5 4.2 4.2 2003 1.8 1.9 Ghana 3.0 3.1 3.1 2003 2.7 1.6 Mali 1.7 1.9 1.9 2001 2.0 1.3 Sources: Initial 2001 estimate from UNAIDS 2002; revised 2001 estimate and 2003 estimate from UNAIDS 2004. as nonresponse, for which adjustments are attempted. The we draw on projections recently prepared for the Central survey reports suggest one important reason why earlier Statistical Offices (CSOs) in Zambia and Botswana (Bulatao UNAIDS estimates could have been too high. In four out 2003a, 2003b). The Zambia projection adopts HIV preva- of five cases, male prevalence is clearly lower than female lence estimates from the 2001­02 DHS rather than from prevalence. The exception, in Burkina Faso, occurs at a low UNAIDS. The DHS data suggest that HIV prevalence among prevalence level. This is in keeping with the typical pattern all adult women is only 75 percent of that among pregnant expected in primarily heterosexual epidemics, where male women, and that male prevalence is still lower. These infections are initially more frequent than female infections, assumptions are adopted for the neighboring country of but the reverse eventually becomes the case. Given that Botswana, for which DHS data are not available. Botswana, UNAIDS applies rates for pregnant women to the popula- with a relatively good health system for the region, has mon- tion as a whole, earlier overestimates are certainly plausible, itored the proportion of males infected, based on clients particularly at higher HIV prevalence levels. Further down- at counseling and testing centers. The proportion of males ward adjustments in prevalence may be needed as more infected is, indeed, lower than that for women. Surveillance surveys become available. They will, of course, come too late data on the age distribution of both men and women for the current set of population projections, all of which-- infected in Botswana also match quite closely the distribu- if this interpretation is correct--start out from too-high tions in the Zambia DHS.3 levels of HIV prevalence and therefore AIDS deaths. Other procedures for these CSO projections follow World Bank methodology. However, in place of WHO life tables incorporating AIDS mortality, Coale-Demeny (1983) Revised Projections life tables without AIDS are used, so that AIDS mortality can These projections may be revised, in later rounds, to allow be specifically modeled. The modeling follows the usual for slightly lower current adult mortality overall and, partic- procedures, involving calculations of HIV incidence from ularly in the most affected countries, either lower or declin- prevalence and estimation of subsequent deaths, relying ing HIV prevalence levels, if further research confirms either on procedures previously described (Bulatao and Bos 1992) to be the case. Pending such revision, precise conclusions but with curves fit to prevalence trends rather than specific about the demographic impact of AIDS in the region cannot calculations, from sexual behavior, of transmission safely be drawn. The revisions to the projections might have probabilities. to be substantial in some cases. For Kenya, for instance, the The Zambia projection begins with a 1980 census, in UNAIDS estimate of HIV prevalence for 2001 was 15.0, order to incorporate the effects of an epidemic initially rec- and the current set of projections depend on this estimate. ognized about 1984, and progresses to 2025. The Botswana The 2003 estimate fell to 6.7, consistent with DHS results. projection begins with a 1981 census and goes to 2031. The The 2001 estimate implies that the maximum reduction in assumed total fertility trends were constructed from census life expectancy--relative to the no-AIDS scenario--would and survey data, some of them only locally available, but are be 17.7 years, using the equation in note 2. The 2003 estimate generally in the range of those that the UN, the Census implies a maximum reduction half as large, at 8.7 years. Bureau, and the World Bank assume for the country. To get a further, although only preliminary, idea of how Figure 6.13 compares the projected population with the projections might change in the most affected countries, previous projections. The effect of AIDS still seems to be Population and Mortality after AIDS | 71 Figure 6.13 Alternative Population Projections, Zambia and Figure 6.14 Alternative Projections of Life Expectancy in Botswana Zambia and Botswana (thousands) Zambia Zambia 80 35,000 70 30,000 60 50 25,000 40 people 20,000 expectancy of 30 life no. 15,000 20 10 10,000 0 1990 2000 2010 2020 2030 2040 2050 5,000 1990 2000 2010 2020 2030 2040 2050 year year no AIDS UN U.S. Census no AIDS UN U.S. Census World Bank Zambia CSO World Bank Zambia CSO Botswana Botswana 80 4,000 70 3,500 60 50 3,000 40 people 2,500 expectancy of 30 life no. 2,000 20 10 1,500 0 1990 2000 2010 2020 2030 2040 2050 1,000 1990 2000 2010 2020 2030 2040 2050 year year no AIDS UN U.S. Census no AIDS UN U.S. Census World Bank Botswana CSO World Bank Botswana CSO Sources: Stanecki 2004; United Nations 2004; World Bank 2004; Bulatao 2003a, 2003b. Sources: Stanecki 2004; United Nations 2004; World Bank 2004; Bulatao 2003a, 2003b. If we look instead at projected life expectancy (fig- large, relative to the UN no-AIDS scenario, but for both ure 6.14), the differences from the other projections stand countries, these CSO projections give higher population out as sharply. For Zambia, the CSO projection shows figures than those from the international agencies. Which levels 5 to 10 years higher than the other projections for the of the preceding projections the CSO projections are closest period 2000­25, although still about 10 years short of what to is determined more by fertility assumptions than the the no-AIDS scenario shows. For Botswana, all three agency modeling of the AIDS effect, since the World Bank has the projections show a similar decline up to about 2000, highest fertility trend for Botswana and the Census Bureau followed by rapid divergence. The CSO projection shows no for Zambia. By 2025, the CSO projections give a Zambia further decline beyond 2001­06, so that about 2025 life population 6 to 32 percent larger than the other projections, expectancy is 15 to 25 years higher than in the other projec- but still almost 20 percent below that in the no-AIDS sce- tions, although almost 15 years below the no-AIDS scenario. nario. For Botswana at the same date, the CSO projection is Adjusting projections to incorporate somewhat lower sur- 10 to 30 percent higher than the others but almost 30 per- vey estimates of HIV prevalence could therefore have a sub- cent below the no-AIDS scenario. Like the World Bank, but stantial effect. In these severely affected countries, it reduces unlike the UN and the Census Bureau, the CSO projections the deficit in life expectancy for 2025 by anywhere from one- show no population decline. third to two-thirds. 72 | Rodolfo A. Bulatao CONCLUSION are strongly affected by HIV/AIDS. However, even in these cases, AIDS is not necessarily the dominant factor in future The demographic impact of AIDS in Sub-Saharan Africa is population trends nor the main source of uncertainty about substantial, but its precise dimensions remain largely a mat- them. Fertility is at least as important. Differing judgments ter of conjecture. Sophisticated models of the course of about future fertility trends can produce at least as great HIV/AIDS epidemics have been developed and linked with variation in future population as the variation between population projection approaches that have proved relative- situations with and without an AIDS epidemic. ly reliable in the past (National Research Council 2000). Fertility itself is affected by HIV/AIDS. Younger women However, these projection models have used inadequate are more likely to be infected than older women, but their data on mortality and HIV/AIDS. eventual deaths leave gaps in the age structure that persist The regional projections, part of the global work done by into and become increasingly evident at older ages. Crude three agencies, the UN Population Division, the U.S. Census birth rates actually rise slightly as a result, because younger, Bureau, and the World Bank, show the population of Sub- more fertile women continue to be replenished in the age Saharan Africa falling below what it would be without AIDS, structure. Unfortunately, this compositional effect is the by 11 to 13 percent by 2020 and by anywhere from 10 to only effect on fertility that can be seen in current projec- 25 percent by 2050. For individual countries, the population tions. Biological and behavioral mechanisms that may lead shortfall is projected as high as 34 to 48 percent by 2020 and to lower fertility are insufficiently understood to be modeled 48 to 68 percent by 2050. That there will be a shortfall rela- across countries. tive to what it would have been had HIV/AIDS not invaded The effect of HIV/AIDS on mortality, the primary mech- the region is certain. That it will be of these dimensions anism by which it affects population growth in agency pro- is not. Each of the projections of countries affected by jections, appears substantial but may be overblown. Life HIV/AIDS starts with UNAIDS estimates for HIV preva- expectancy is projected to fall a maximum of one year below lence in 2001, some of which were clearly overestimates. its expected path for every one percentage point increase For how many this is true cannot yet be told. in initial HIV prevalence. With UNAIDS adjusting its esti- Between 2002 and 2004, UNAIDS generally revised mates of HIV prevalence downward, life expectancy should downward its estimates for 2001 adult HIV prevalence in probably be expected not to decline as much. For Kenya, for Sub-Saharan Africa. DHS data show sharply lower estimates example, UNAIDS reduced its estimate of 2001 HIV preva- of adult prevalence in three countries, and UNAIDS made lence from 15 to 8 percent. The projections show maximum matching adjustments for two (and somewhat smaller loss in years of life ranging from 15 to 21 years, but a small- adjustments for the rest). For countries where HIV preva- er loss of 8 to 14 years would seem more likely. lence is high, initial overestimates could have been due to Given that the HIV/AIDS epidemics in Sub-Saharan the practice of generalizing prevalence estimates for preg- Africa are mainly heterosexual, the rise in mortality affects nant women to the adult population. Prevalence among women more than men. The female advantage in life males tends to be lower than among females, particularly at expectancy will shrink and could turn into a female disad- higher levels. vantage. This effect too depends on levels of HIV preva- Unfortunately, the population projections available rely lence, and the projections may overstate the change. Effects on the earlier estimates of HIV prevalence. This could lead on the sex ratio, as well as on age structure, are more to overstated base adult mortality, as well as to an overstated substantial at high HIV prevalence levels, but even in these future trend in AIDS deaths. Comparing projections that cases they do not appear to produce patterns more extreme rely on survey-based HIV estimates suggests that, in severely than those seen in other populations. affected countries, the reduction in population due to AIDS Each agency projection will undoubtedly be revised in may be only 50 to 75 percent of what the agency projections the future to take adjusted HIV prevalence levels into show and that life expectancy may be reduced, at the account. How they came to adopt HIV prevalence levels that maximum, by a third to two-thirds of what agency projec- were too high is an interesting question. The agencies devote tions show. some effort to ensuring that they use the best possible esti- If the agency projections are not, at present, a reliable mates of current fertility and mortality. Perhaps they have guide to future populations in areas affected by HIV/AIDS, not spent as much effort carefully considering the accuracy they nevertheless illustrate patterns of demographic impact. of input data on HIV/AIDS and have been insufficiently The effect on population does appear larger when countries critical of UNAIDS reports. Even with the adjustments Population and Mortality after AIDS | 73 made to previous HIV prevalence estimates, there remain ------. 2003b. "Population Projections for Zambia, 2000­2025." Paper prepared for the Central Statistical Office, Lusaka, Zambia. some that probably deserve closer scrutiny. Bulatao, R. A., and E. Bos. 1992. "Projecting the Demographic Impact There also remain other challenges to future rounds of of AIDS." Policy Research Working Paper 941, World Bank, these projections. Although data on HIV prevalence are Washington, DC. gradually improving, what path the epidemics will take on Coale, A. J., and P. Demeny, with B. Vaughn. 1983. Regional Model Life Tables and Stable Populations. 2nd ed. New York: Academic Press. the downslope remains conjectural. How therapies and vac- Lopez, A. D., O. B. Ahman, M. Guillot, B. D. Ferguson, J. A. Salomon, and cines may help, what behavior changes can be anticipated, C. J. L. Murray. 2002. World Mortality in 2000: Life Tables for 191 and what the implications for infections and deaths will be Countries. Geneva: WHO. are matters about which little if any empirical information Lopez, A. D., J. Salomon, O. Ahman, C. J. L. Murray, and D. Mafat. 2001. "Life Tables for 191 Countries: Data, Methods and Results." Global can be adduced. In this area projections are, at best, edu- Program on Evidence for Health Policy, Discussion Paper 9, WHO, cated guesses. Geneva. Another challenge will be managing the complexity of McDevitt, T. M. 1999. World Population Profile: 1998. Report WP/98. the methodology to project AIDS mortality. Complexity Washington, DC: U.S. Census Bureau. Murray, C. J. L., O. B. Ahmad, A. D. Lopez, and J. Salomon. 2000. "WHO need not itself be a problem, but where it disguises a poverty System of Model Life Tables." Global Program on Evidence for Health of data and lends an inappropriate air of authoritativeness Policy, Discussion Paper 8, WHO, Geneva. to results while concealing the calculations on which they National Research Council. 2000. Beyond Six Billion: Forecasting the are based, it can easily mislead. Projections need to convince World's Population. Washington, DC: National Academy Press. Salomon, J. A., E. E. Gakidou, and C. J. L. Murray. 1999. "Methods for not only with sophisticated models but also with a good Modelling the HIV/AIDS Epidemic in Sub-Saharan Africa." Global appreciation of historical patterns and trends. More trans- Program on Evidence for Health Policy, Discussion Paper 3, WHO, parency would probably make it easier not just to detect the Geneva. weaknesses of the projections but also to better appreciate Stanecki, K. A. 2004. The AIDS Pandemic in the 21st Century. U.S. Census Bureau, International Population Reports WP02-2. Washington, DC: what they actually contribute. U.S. Government Printing Office. Stanley, E. A., S. T. Seitz, P. O. Way, P. D. Johnson, and T. F. Curry. 1991. "The IwgAIDS Model for the Heterosexual Spread of HIV and the Demographic Impacts of the AIDS Epidemic." In The AIDS Epidemic NOTES and Its Demographic Consequences. New York: United Nations. Timaeus, I. M., and M. Jasseh. 2004. "Adult Mortality in Sub-Saharan 1. If Pt is the projected population to time t and Pt* is the projected Africa: Evidence from Demographic and Health Surveys." Demography population with which it is being compared, then Pt Pt* P0 P0* e(b­b*)t 41 (4): 757­72. e( d d*)te( m m*)t , where P0 is the base population for the projection and UNAIDS. 2000. Report on the Global HIV/AIDS Epidemic­June 2000. b, d, and m represent crude birth, death, and net migration rates. Bulatao Geneva: UNAIDS. (2001) uses a similar decomposition in assessing projection accuracy. 2. The equation, derived from the UN projection, is: Loss in e0 ------. 2002. Report on the Global HIV/AIDS Epidemic 2002. Geneva: 1.5 1.08 HIV adult prevalence in percent in 2001 (R2 0.97). UNAIDS. 3. Preliminary results from the 2004 Botswana AIDS Impact Survey II, ------. 2004. Report on the Global AIDS Epidemic, July 2004. Geneva: conducted after these projections were completed, appear to confirm the UNAIDS. need to use substantially lower adult HIV prevalence levels in projections. UNAIDS Reference Group on Estimates, Modelling and Projections. 2002. They show adult HIV prevalence at 25.3 percent, well below the 37.3 per- "Improved Methods and Assumptions for Estimation of the HIV/ cent that UNAIDS reports. AIDS Epidemic and Its Impact: Recommendations of the UNAIDS Reference Group on Estimates, Modelling and Projections." AIDS 16: W1­14. REFERENCES United Nations. 1998. World Population Prospects: The 1998 Revision. New York: United Nations. Ahmad, O. B., A. D. Lopez, and M. Inoue. 2000. "The Decline in Trends in ------. 2003a. World Population Prospects: The 2002 Revision. Vol. 1: Child Mortality: A Reappraisal." Bulletin of the World Health Comprehensive Tables. New York: United Nations. Organization 78 (10): 1175­91. ------. 2003b. World Population Prospects: The 2002 Revision. Vol. 2: Sex Bos, E., M. T. Vu, E. Massiah, and R. A. Bulatao. 1994. World Population and Age Distribution of the World Population. New York: United Projections, 1994­95 Edition: Estimates and Projections with Related Nations. Demographic Statistics. Baltimore: Johns Hopkins University Press. ------. 2004. World Population Prospects: The 2002 Revision. Vol. 3: Bulatao, R. A. 2001. "Visible and Invisible Sources of Errors in World Analytical Report. New York: United Nations. Population Projections." Paper presented at the annual meeting of the World Bank. 2004. World Development Indicators 2004. Washington, DC: Population Association of America, Washington, DC, March. World Bank. ------. 2003a. "Population Projections for Botswana up to 2031." Paper Zaba, B., and S. Gregson. 1998. "Measuring the Impact of HIV on Fertility prepared for the Central Statistical Office, Gaborone, Botswana. in Africa." AIDS 12 (Suppl. 1): S41­50. 74 | Rodolfo A. Bulatao Chapter 7 Levels and Patterns of Mortality at INDEPTH Demographic Surveillance Systems Osman A. Sankoh, Pierre Ngom, Samuel J. Clark, Don de Savigny, and Fred Binka Empirical, longitudinal, population-based data on mortality established for these trials have continued to operate long in Africa have, until recently, been unavailable. This critical after the end of the original trials and have proved increas- information gap for Sub-Saharan Africa became even more ingly useful for both researchers and policy makers evident as an impediment to our understanding of health (Armstrong et al. 1999; de Savigny et al. 1999; Tanzania and disease in Africa with the arrival of a major new cause Ministry of Health 1997). of rapidly increasing mortality in the form of HIV/AIDS. Accurate data on mortality conditions in Africa are still At about the same time, during the 1990s, many large-scale scarce. Until recently, the main tool for bridging this gap was mortality intervention trials were conducted at the commu- the use of indirect demographic estimation techniques and nity level, principally for understanding the efficacy of new model age-specific mortality schedules produced by Brass interventions such as vitamin A supplementation, and the and colleagues (1973), Coale and Demeny (1966), and the use of insecticide-treated mosquito nets for malaria control United Nations Department of International Economic and (Alonso et al. 1991; Binka et al. 1996; Nevill et al. 1996; Ross Social Affairs (1982). The Brass relational system is based on et al. 1995). All these trials used demographic surveillance empirical data collected in West Africa during the middle of systems (DSSs) to measure the impact of mortality. These the twentieth century. In contrast, neither the Coale and successful mortality intervention trial efforts focused Demeny nor the UN model life table systems use significant renewed attention on the usefulness of DSSs for illumi- amounts of data collected from Africa. Moreover, all three of nating the fundamental age, sex, and cause structure of the systems are based on 30- to 50-year-old data. Given the mortality in resource-constrained settings in Africa. The dramatic demographic changes that have affected Africa in malaria intervention trials themselves forced a complete the past 20 to 30 years and the fact that two of the systems reconsideration of the previously underestimated role of are based largely on data collected from other regions of malaria as a direct and indirect cause of mortality in Africa the world, whereas the third is based on data from only one (Breman, Egan, and Keusch 2002). More and more DSS sites region of Africa, it may be problematic to use them in the 75 current African context. No doubt, the World Fertility supervisors, larger numbers of community key-informants Survey (WFS) and the Demographic and Health Surveys continuously notify the system of birth or death events for (DHSs) have remedied the above situation in part by immediate follow-up. Only events occurring to registered increasing our knowledge of the level, trends, and differen- households and members are linked to the database for tials in infant and child mortality in the developing world. analysis. Sophisticated field operational logistics and linked Also, since independence, several African countries have data quality control and computing systems support this undertaken national censuses, but mortality data from these routine DSS surveillance (MacLeod, Phillips, and Binka sources are often plagued with underreporting and need, 1995; Phillips, MacLeod, and Pence 2000). DSS sites can therefore, to be adjusted using hypotheses that are not generate accurate population structures, person-years at always realistic. risk, mortality rates, cause-specific mortality rates, propor- Much of the developing world is not adequately covered tional mortality, fertility rates, and rates of internal and by accurate vital registration systems, leading to a substan- external migrations. In addition, because of the intensity of tial lack of direct empirical data describing mortality. household survey frequency, DSS sites commonly docu- Because planning must still be undertaken, information ment a large array of contextual information on household on mortality is inferred through interpolation or extrapo- structures, dependency, occupation, education, access to lation from existing commonly observed age patterns of and use of services, as well as socioeconomic status. A mortality--the so-called model life tables. These model typical DSS produces more than 100 health- and poverty- age patterns of mortality are used to substitute, extend, or monitoring indicators annually on each household. Details fill in where observed mortality data are missing and are of DSS concepts and methods are provided by INDEPTH often key ingredients in methods used to estimate demo- Network (2002). graphic indicators from sparse data. Underlying epi- On the completeness and reliability of data produced by demiological profiles are inferred based on those that are DSS, Korenromp and colleagues (2003) wrote in their known to underlie the existing model of mortality patterns review paper, "For Africa, with an increasing number of sites that most closely match the observed data. The closest- that monitor cause-specific mortality at a population level fitting model patterns are then used, among other things, by means of standard methods under the INDEPTH net- to estimate levels of child mortality and create population work, DSS may at present form one of the more complete projections. data sources." And WHO/UNICEF (2003, 20) conclude in DSSs bridge the data gap that exists in resource- their Africa Malaria Report--2003 that "[a]t present, the constrained countries. A DSS refers to a methodological most reliable data on trends in malaria death in children approach for monitoring a registered, dynamic cohort of the under 5 years of age is obtained from demographic surveil- total population of a geographically confined area. Typical lance systems (DSS)." DSS populations monitored include at least 60,000 individ- In 1998 a large number of DSS sites formed an interna- uals per site. This is usually sufficient to provide adequate tional network called the INDEPTH Network. Details of sample sizes to monitor trends in all-cause mortality and the sites, their locations, the populations they follow, and the most common cause-specific mortality, by sex and age basic demographic outputs of each site are available on the group. Members of the dynamic DSS cohort are registered Internet (www.indepth-network.net) and in INDEPTH as such during an initial census. New members enter the Network (2002). The central purpose of the INDEPTH cohort either by system-registered births or in-migrations, Network is to facilitate cross-site analyses of comparable and members exit either by system-registered deaths or data across broader geographic areas with more diverse cir- by out-migration. Causes of death are determined for all cumstances in order to answer questions that cannot be deaths by verbal autopsy. Continuous cycles of reenumera- answered within individual sites. Comparative mortality tion maintain an accurate person-time denominator and and mortality patterns are an obvious first endeavor of such help identify numerator events such as pregnancies, births, a network, and these are provided in this chapter. By the deaths, and migration. These cycles of reenumeration typi- efforts of the members of the INDEPTH Network, for the cally take place three to four times per year in order to opti- first time schedules of mortality in Africa and empirical mize the chance to identify pregnancies and thus determine data on age standardization and on age and sex structure outcomes of pregnancy. In addition to the routine data are available. This network is the source of the analyses capture performed by specialist teams of enumerators and presented in this chapter. 76 | Osman A. Sankoh, Pierre Ngom, Samuel J. Clark, Don de Savigny, and Fred Binka Standardized mortality rates were computed based on deaths occurred. An average of about 17 percent of the the new INDEPTH Network standard population that for person-years exposed were lived at ages younger than five the first time is derived from empirical data. The INDEPTH years, and an average of 39 percent of the deaths also standard population typifies the true structure of the young occurred between birth and age five. The crude death rate population in Sub-Saharan Africa. The chapter then pres- for both sexes combined ranges from a low of 7 per 1,000 in ents basic life table indicators for INDEPTH sites, based on Agincourt, South Africa, to 39 per 1,000 in Bandim, Guinea- their age-specific mortality rates over the 1995­99 period. Bissau. Seven new mortality patterns are developed from more than 4.2 million person-years of observation at the African Overall Mortality INDEPTH sites. These patterns are demonstrated to be sub- stantially different from conventionally used model mortal- The crude death rate and the expectation of life at birth are ity patterns applied to Africa. the two indicators used in this section to examine overall levels of mortality at the INDEPTH sites. In order to remove the influence of each site's age structure and to make the LEVELS OF MORTALITY AT INDEPTH comparison of the crude death rates more reasonable, it is NETWORK SITES necessary to standardize such rates. To do this, there are several widely used standard age distributions, including The data used in this chapter come from 17 DSS sites in Segi's world standard and World Health Organization Sub-Saharan Africa for which information on mortality was (WHO) standard age distributions (Estève, Benhamou, and available for at least a full year from 1995 to 1999 (table 7.1). Raymond 1994). Both of these standards reflect populations The overall average length of the observation period for with relatively low fertility and mortality. Consequently, the contributing sites is four years. The data yield a total of they give significant weight to the middle years of life. All the 3,979,155 person-years of exposure, during which 55,356 INDEPTH sites record information from relatively young Table 7.1 Summary of Mortality Data from INDEPTH Sites, 1995­99 Reporting Duration Observed Observed Crude Deaths Person-years Site Country period of period deaths person-years death rate under age 5 (%) under age 5 (%) Agincourt South Africa 1995­99 5 1,738 304,530 7.11 15.54 13.79 Bandafassi Senegal 1995­99 5 901 41,286 33.57 53.16 19.86 Bandim Guinea-Bissau 1995­97 3 1,830 64,434 38.65 56.01 27.69 Butajira Ethiopia 1995­96 2 834 72,873 19.20 41.49 16.94 Dar es Salaam Tanzania 1994/95­1998/99a 5 4,515 354,041 21.75 27.44 13.87 Farefenni The Gambia 1995­99 5 1,201 81,872 21.23 45.05 17.12 Gwembe Zambia 1991­95 5 576 37,089 26.89 59.72 19.37 Hai Tanzania 1994/95­1998/99a 5 8,106 746,864 16.09 23.14 14.30 Ifakara Tanzania 1997­99 3 1,812 159,639 20.28 41.17 16.23 Manhica Mozambique 1998­99 2 973 67,344 20.97 35.66 17.06 Mlomp Senegal 1995­99 5 374 37,051 13.75 20.59 10.80 Morogoro Tanzania 1994/95­1998/99a 5 9,548 538,286 30.01 29.03 13.01 Navrongo Ghana 1995­99 5 11,278 691,679 27.72 34.46 14.10 Niakhar Senegal 1995­98 4 1,993 116,133 24.30 51.03 18.05 Nouna Burkina Faso 1995­98 4 1,650 117,156 17.00 40.48 18.24 Oubritenga Burkina Faso 1995­98 4 6,967 478,315 24.83 49.63 17.40 Rufiji Tanzania 1999 1 1,060 70,563 33.96 35.47 16.32 Total or average 4 55,356 3,979,155 23.37 38.77 16.71 Source: INDEPTH Network 2002. a. Reporting in mid-year to mid-year annual periods resulted in a five-year reporting period from July 15, 1994, to July 15, 1998. Levels and Patterns of Mortality at INDEPTH Demographic Surveillance Systems | 77 populations with comparatively high fertility and high Table 7.2 displays the crude death rate for each site and mortality. Under those conditions, there are proportionally the age-standardized crude death rates calculated using both more young people in the population, giving it a "younger" the INDEPTH and Segi standard age distributions along age distribution. When the Segi or WHO standard age dis- with the values for the expectation of life at birth. tributions are applied to the INDEPTH data, they give too The INDEPTH Network standardized crude death rates much weight to the relatively high mortality rates prevailing range from 7 to 33 per 1,000 for males and 5 to 27 per 1,000 at middle and older ages and too little weight to mortality at for females, revealing a wide range of mortality at the younger ages. Consequently, the absolute level of the age- INDEPTH sites. The figures for the expectation of life at standardized crude death rates produced with those stan- birth vary in a relationship that is loosely inverse to the dards significantly overestimates the true level of mortality values of the crude death rate (figure 7.2), and they cover a at the INDEPTH Network sites. similarly wide range of from 66 to 36 years for males and 74 To address this problem and create age-standardized to 40 years for females. The data from Bandim are anom- crude death rates that more accurately reflect the true level alous and reflect some unresolved questions concerning the of mortality at the INDEPTH sites, we have constructed manner in which they were collected and reported. the INDEPTH standard age distribution. An average age There is some geographic clustering. Together at the low distribution is constructed for each site over the period end of the spectrum are three rural sites from Tanzania 1995­99 by taking the weighted average of the person-years (Hai, Ifakara, and Rufiji) and one site in Senegal (Mlomp). of exposure in each age group across all the years for which In the middle of the pack are three sites in West Africa: data are reported. The weight for each year is the total Farefenni, Nouna, and Oubritenga. At the high end there is number of person-years reported for all ages during that a mixture of sites from West, East, and southern Africa. The year. The INDEPTH standard age distribution is calculated absolute level of mortality varies considerably over space; by taking the weighted average of the individual site aver- sites located close to each other have similar levels of mor- age age distributions in each age group. In this case the tality, but all major regions of Africa show a wide range of weights are the total number of person-years in each of mortality levels. the individual site average age distributions. In figure 7.1, For the most part the sex differentials are relatively small the younger age distribution of the INDEPTH standard, but are generally in favor of females, as expected. Two of the which is typical of developing countries, is contrasted with sites in southern Africa--Agincourt, in South Africa, and the much older population structures of the Segi and WHO Manhica, in Mozambique--have significant male migration standards. and register more substantial sex differentials, which stand out in contrast to the rest of the sites. Bandim, West Africa, also records a substantial sex differential, but as noted above Figure 7.1 Standard Population Age Structure from INDEPTH, Segi, and WHO there may be a methodological explanation for this. 0.16 0.14 Infant and Child Mortality 0.12 The measures of child mortality displayed in table 7.3 are the life table probabilities of dying in a specified age group: 0.1 1 0 q for ages zero to one year, q for ages one to five years, population 4 1 0.08 and q for ages zero to five years. The conventional infant total 5 0 of 0.06 mortality rate--the ratio of the number of deaths below one % 0.04 year of age and the number of births for a given period--is also included. 0.02 As shown in figure 7.3, there is a wide range in the level 0 of child mortality. The probability that a newborn dies 0­4 5­9 10­1415­19 20­2425­29 30­3435­39 40­4445­49 50­5455­59 60­6465­69 70­7475­79 80­8485 age before reaching its fifth birthday ranges from 32 per 1,000 to 255 per 1,000 for males and 34 per 1,000 to 217 per 1,000 for INDEPTH Segi WHO females. The Agincourt site in South Africa records a Source: INDEPTH Network 2002. comparatively low level of child mortality. Another cluster, 78 | Osman A. Sankoh, Pierre Ngom, Samuel J. Clark, Don de Savigny, and Fred Binka Table 7.2 Crude Death Rates and Expectation of Life at Birth (per thousand) Male Female Site CDR ASCDRI ASCDRS e0 CDR ASCDRI ASCDRS e0 Agincourt 5.93 7.42 9.43 66.12 4.65 4.90 5.90 74.38 Mlomp 10.35 10.80 12.51 60.46 9.83 8.59 9.68 64.78 Hai 12.33 11.56 13.49 56.26 9.49 8.65 9.74 62.80 Rufiji 14.67 12.19 13.57 53.40 15.35 12.61 13.28 52.18 Ifakara 11.70 12.45 13.98 55.73 11.01 11.37 12.28 58.22 Butajira 11.65 12.50 13.79 55.81 11.25 12.44 13.50 56.68 Nouna 13.74 13.62 14.46 54.20 14.42 14.41 15.71 53.06 Oubritenga 15.68 14.93 15.95 51.63 13.58 13.05 13.53 55.08 Farefenni 16.24 15.84 17.47 50.83 13.17 13.56 14.08 55.05 Dar es Salaam 12.84 17.15 20.52 50.32 12.66 16.45 19.42 49.76 Niakhar 18.45 17.45 18.26 48.80 15.89 14.40 14.81 53.59 Manhica 17.00 17.50 20.11 47.47 12.41 11.36 12.60 58.12 Navrongo 17.66 18.07 20.42 47.22 15.10 15.82 17.66 51.39 Gwembe 18.69 19.27 21.89 47.32 16.82 17.95 19.67 53.66 Morogoro 18.70 19.27 21.90 44.44 16.82 17.95 19.67 46.11 Bandafassi 23.49 20.62 21.62 44.74 20.36 18.30 18.71 47.54 Bandim 31.35 32.86 38.63 35.86 25.65 27.48 31.42 38.91 Source: INDEPTH Network 2002. Note: CDR crude death rate; ASCDR age-standardized crude death rate; I standardized with INDEPTH Standard Age Structure; S standardized with Segi Standard Age Structure; e0 expectation of life at birth. composed of Mlomp in Senegal and Hai in Tanzania, reports low levels of child mortality but not nearly as low as that of the South Africa site. The next higher cluster is composed of Figure 7.2 Crude Death Rate and Expectation of Life at Birth sites from various regions of Africa including Dar es Salaam, 35 80 Tanzania; Butajira, Ethiopia; Ifakara, Tanzania; Nouna, Burkina Faso; and Manhica, Mozambique. Following after 30 70 that with 5q0 close to 175 per 1,000 for males and females are 60 25 Farefenni, The Gambia; Rufiji, Tanzania; Navrongo, Ghana; 50 Gwembe, Zambia; Morogoro, Tanzania; and Oubritenga, 1 20 40 Burkina Faso. The three remaining sites--Niakhar, Senegal; (years) ASCDR 15 e0 Bandim, Guinea-Bissau; and Bandafassi, Senegal--all have 30 10 substantially higher values of 5 0 q closer to 225 per 1,000. 20 There is a wide range in the level of child mortality, but 5 10 except at the lowest and highest levels, there does not appear 0 0 to be any geographical clustering. The lowest levels are defi- p ji ira ca be si m AgincourtMl om Hai Rufi ara Ifak Butaj Nouna tenga rongo em Bandi nitely found in South Africa, whereas the highest levels are Oubri FarefenniSalaamNiakhar es ManhiNav Gw Morogoro Dar Bandafas reported from West Africa. INDEPTH sites Table 7.3 also displays the ratio of q 1 0 to q in order to 4 1 Male ASCDR1 Female ASCDR1 Male e0 Female e0 elucidate the changing risk of death faced by children before and after their first birthday.1 This ratio reveals that children Source: INDEPTH Network 2002. in Rufiji who survive to age one face a probability of death Note: ASCDR1 age-standardized crude death rate, INDEPTH Standard Age Structure; e0 expectation of life at birth. that is improved by nearly a factor of four, whereas children Levels and Patterns of Mortality at INDEPTH Demographic Surveillance Systems | 79 Table 7.3 Infant and Child Mortality (per thousand) Male Female Site IMR 1 0 q 4 1 q 5 0 q 1 0 4 1 q / q 1 0 q 4 1 q 5 0 q 1 0 4 1 q / q Agincourt 16.93 15.06 17.52 32.32 0.86 16.63 17.35 33.69 0.96 Mlomp 45.18 48.24 42.61 88.80 1.13 49.42 51.74 98.60 0.96 Hai 67.13 66.78 26.73 91.73 2.50 56.54 26.68 81.71 2.12 Dar es Salaam 71.13 66.38 50.86 113.86 1.30 67.20 52.49 116.16 1.28 Butajira 67.82 65.62 57.73 119.56 1.14 71.09 62.20 128.87 1.14 Ifakara 93.22 76.12 52.23 124.37 1.46 86.09 50.27 132.03 1.71 Nouna 40.85 34.31 107.53 138.15 0.32 42.71 106.82 144.97 0.40 Manhica 72.65 85.75 68.91 148.75 1.24 59.37 60.41 116.19 0.98 Farefenni 74.65 68.04 110.47 171.00 0.62 66.46 109.12 168.32 0.61 Rufiji -- 91.0 46.4 133.2 1.96 110.0 35.1 141.6 3.13 Navrongo 109.59 106.58 83.54 181.21 1.28 102.96 73.23 168.65 1.41 Gwembe -- 105.24 87.26 183.32 1.21 111.94 78.78 181.90 1.42 Morogoro 116.73 105.24 87.26 183.32 1.21 111.94 78.78 181.90 1.42 Oubritenga 96.49 102.25 95.97 188.41 1.07 91.88 104.84 187.09 0.88 Niakhar -- 89.80 146.84 223.45 0.61 72.16 129.14 191.98 0.56 Bandim -- 112.37 129.78 227.57 0.87 101.52 128.31 216.80 0.79 Bandafassi 124.88 138.60 134.59 254.54 1.03 116.43 114.29 217.42 1.02 Source: INDEPTH Network 2002. Note: IMR infant mortality ratio; -- not available. Figure 7.3 Child Mortality: Probability of Dying between Birth Senegal; and Bandafassi, Senegal--where there is a clear dif- and Age Five (5q0) ference favoring females, except in Rufiji, where males are 300 favored. 250 200 Adult Mortality 1,000 150 per Values for the probability that persons who survive to their 0 q twentieth birthday will survive to their fiftieth birthday, that 5 100 50 is,30 20 q , are displayed in table 7.4 along with values of q 5 0 and the ratio of q to q .2 There are wide variations in 0 5 0 30 20 incourtMlo mp i o rogorbritengaNia o Ha una fiji be the levels of adult mortality from 159 per 1,000 to 501 per Ru khar ndim Ag Dar es SalaamButajiraIfakara No Manhicaefenni Far Ba Na vrongGwemMo Ou ndafassi Ba 1,000 for males and 112 per 1,000 to 421 per 1,000 for INDEPTH sites females. A number of sites record substantial differences in adult mortality between the sexes--Agincourt, South Africa; male female Hai, Tanzania; Manhica, Mozambique; Mlomp, Senegal; and Source: INDEPTH Network 2002. Navrongo, Ghana, in particular. There also exists the opposite differential, wherein female mortality rates exceed in Bandafassi face a nearly constant probability of dying male rates, at two sites: Rufiji, Tanzania; and Dar es Salaam, throughout the first five years of life. Tanzania. The human immunodeficiency virus and Differences in child mortality according to sex are rela- acquired immune deficiency syndrome (HIV/AIDS) and tively small and do not appear to consistently favor one sex maternal mortality may explain these patterns. over the other. Interestingly this pattern is broken by four For the first time, Agincourt in South Africa does not sites--Manhica, Mozambique; Rufiji, Tanzania; Niakhar, define the low end of the range. There also does not appear 80 | Osman A. Sankoh, Pierre Ngom, Samuel J. Clark, Don de Savigny, and Fred Binka Table 7.4 Adult Mortality and Child and Adult Mortality Ratio (per thousand) Male Female Site 5 0 q 30 20 q 5 0 30 20 q / q 5 0 q 30 20 q 5 0 30 20 q / q Mlomp 88.80 159.03 0.56 98.60 111.51 0.88 Niakhar 223.45 165.25 1.35 191.98 141.86 1.35 Agincourt 32.32 196.35 0.16 33.69 100.77 0.33 Nouna 138.15 199.93 0.69 144.97 184.51 0.79 Farefenni 171.00 205.13 0.83 168.32 149.88 1.12 Oubritenga 188.41 210.62 0.89 187.09 157.60 1.19 Bandafassi 254.54 226.27 1.12 217.42 200.42 1.08 Butajira 119.56 227.19 0.53 128.87 193.86 0.66 Rufiji 133.2 234.77 0.57 141.6 268.42 0.53 Ifakara 124.37 240.09 0.52 132.03 185.07 0.71 Navrongo 181.21 298.01 0.61 168.65 188.86 0.89 Hai 91.73 304.77 0.30 81.71 229.38 0.36 Dar es Salaam 113.86 331.46 0.34 116.16 369.74 0.31 Manhica 148.75 382.13 0.39 116.19 197.39 0.59 Gwembe 183.32 408.82 0.45 181.90 372.81 0.49 Morogoro 183.32 409.03 0.45 181.90 372.81 0.49 Bandim 227.57 500.75 0.45 216.80 421.42 0.51 Source: INDEPTH Network 2002. to be any substantial geographical clustering of similar risk the dimensionality of the data to as few dimensions as of adult mortality within the region. The cluster of moder- possible; and the provision of a common reference to which ate risk includes sites from all major regions of Sub-Saharan all of the observed patterns can be compared. Africa as does the high risk cluster. The relationship between child and adult mortality Principal Components reveals two distinct groups: sites in West Africa and those in the rest of Africa. Some of the West African sites clearly The principal components technique is used to identify record levels of child mortality that are high relative to the 15 age components that together are able to represent all but corresponding levels of adult mortality. In the West African a negligible amount of the variation in age patterns of mor- sites of Bandafassi, Senegal; Farefenni, The Gambia; tality among the 70 male and female site periods. In fact, the Oubritenga, Burkina Faso; and Niakhar, Senegal, this is the first four of these components represent the vast majority of result of unusually high child mortality coupled with sub- the variation in the original data. When recombined with stantial adult mortality. the appropriate weights the components are able to repre- sent all the observed age patterns of mortality in the INDEPTH database, and furthermore, by choosing different INDEPTH NETWORK MORTALITY PATTERNS weights any arbitrary age pattern of mortality can be repre- sented to within a negligible error. This section summarizes the method used to identify the The first four principal components address all three of seven INDEPTH patterns of mortality. For a full description the key considerations important to identifying common of the process, see INDEPTH Network 2002, chap. 7. Key underlying patterns: by dropping the other components considerations in identifying common underlying patterns much of the small-magnitude, random noise is eliminated; (empirical regularities) are the filtering out of small, unim- the dimensionality of the data is reduced to four weights portant, and potentially random variation; the reduction of instead of 18 age groups; and the four principal components Levels and Patterns of Mortality at INDEPTH Demographic Surveillance Systems | 81 provide a common reference against which all the observed Pattern 1. The first pattern is similar to the Coale-Demeny patterns can be compared. The task then becomes how to North and UN Latin American model life table age patterns identify similar observed patterns. of mortality (INDEPTH Network 2002). There is no indica- tion that HIV/AIDS affects pattern 1, and the male and female age patterns are similar with the exception of a bulge Clustering in the female pattern during the reproductive years, pre- The aim of cluster analysis is to identify groups of similar sumably caused by maternal mortality. Pattern 1 is primarily objects. The objects are usually described by a vector of derived from sites in West Africa over the entire period cov- numbers representing measurements of some attributes of ered by the INDEPTH Network data set. HIV/AIDS has not the objects. In this case, the vectors each contain four coeffi- yet become as significant a problem in West Africa as it is in cients representing the weights on the first four principal central and southern Africa, so a large impact of AIDS in the components that are necessary to reproduce a given data from West Africa is not expected. It is worth noting that observed mortality pattern.3 The hierarchical clustering child mortality between the ages of one and nine is signifi- algorithm is used to identify five clusters of similar mortality cant and substantially elevated above the most similar exist- patterns for males and seven for females. Because the seven ing models. This is in keeping with the fact that malaria is a female patterns are quite different from one another and significant cause of death in West Africa, and it has a large because the male patterns corresponding to the two addi- impact on those ages. tional female patterns are grouped together in the male clus- ters, the seven female clusters are chosen as the final clusters Pattern 2. The INDEPTH Network pattern 2 is derived of observed mortality patterns. largely from Asian data and for that reason is not discussed Once the clusters are identified the person-year weighted here; see INDEPTH Network 2002 for a discussion of average of the individual age patterns in each cluster is cal- pattern 2. culated to yield the characteristic age pattern of mortality for each cluster. Pattern 3. The sites contributing to pattern 3 are almost exclusively located in southern Africa and East Africa, South Africa and Tanzania in particular. This pattern obviously The Seven INDEPTH Mortality Patterns contains some influence of HIV/AIDS but not nearly to the The seven commonly observed age patterns of mortality degree observed in pattern 5. The South African data come emerging from the INDEPTH data are sufficiently different from the Agincourt site (INDEPTH Network 2002), where from the existing model life tables (INDEPTH Network mortality is extraordinarily low compared with the other 2002) to qualify as new mortality patterns; and because INDEPTH sites in Africa, and where HIV/AIDS is recog- almost all the data on which they are based come from nized but does not yet affect the population in the cata- Africa, they are African patterns appropriate for use in strophic sense that it does in other parts of southern and Africa. The natural log­transformed, ln(nqx), values are dis- East Africa. The remainder of the data come from the Dar es played in figure 7.4. Salaam site, where the impact of HIV/AIDS appears to be Underlying epidemiological profiles underpinning each greater. This pattern is most similar to the UN Far East pat- pattern are still being identified. The INDEPTH Network is tern of mortality, corresponding to the low infant and child actively engaged in gathering and analyzing information on mortality compared with mortality at older ages (INDEPTH cause of death corresponding to the raw mortality rates pre- Network 2002). A noteworthy feature of this pattern is that sented here. Once that work is complete, it will be possible infant and child mortality does not appear to be substan- to identify the specific causes of death contributing to each tially elevated, as might be expected when HIV/AIDS is an pattern. However, until that time we must speculate and important contributor to mortality. Timaeus (1998) sug- infer from what is known about the regional distribution of gests that the strong secular decrease in child mortality that major causes of death in Africa. has been observed in many settings in Africa during the lat- The following discussion of the patterns has largely been ter half of the twentieth century may largely cancel out the excerpted from chapter 7 of Population, Health, and Survival increase in child mortality brought about by HIV to yield at INDEPTH Sites (INDEPTH Network 2002). approximately unchanging child mortality, rather than the 82 | Osman A. Sankoh, Pierre Ngom, Samuel J. Clark, Don de Savigny, and Fred Binka Figure 7.4 INDEPTH Mortality Patterns 1­7, ln(nqx) Pattern 1 Pattern 2 0.5 0.5 0.5 0.5 1.5 1.5 ) ) x x q( n 2.5 q( n 2.5 In In 3.5 3.5 4.5 4.5 5.5 5.5 0 0 1­4 5­9 10­1415­19 1­4 20­2425­29 30­3435­39 40­4445­49 50­5455­59 60­6465­69 70­74 75­7980­84 5­9 10­1415­1920­2425­29 30­3435­39 40­4445­49 50­5455­59 60­6465­69 70­74 75­7980­84 age (years) age (years) Pattern 3 Pattern 4 0.5 0.5 0.5 0.5 1.5 1.5 ) ) x x q( n 2.5 q( n 2.5 In In 3.5 3.5 4.5 4.5 5.5 5.5 0 0 1­4 5­9 10­1415­19 1­4 20­2425­29 30­3435­39 40­4445­49 50­5455­59 60­6465­69 70­74 75­7980­84 5­9 10­1415­1920­2425­29 30­3435­39 40­4445­49 50­5455­59 60­6465­69 70­74 75­7980­84 age (years) age (years) Pattern 5 Pattern 6 0.5 0.5 0.5 0.5 1.5 1.5 ) ) x x q( n 2.5 q( n 2.5 In In 3.5 3.5 4.5 4.5 5.5 5.5 0 0 1­4 5­9 10­1415­19 1­4 20­2425­29 30­3435­39 40­4445­49 50­5455­59 60­6465­69 70­74 75­7980­84 5­9 10­1415­1920­2425­29 30­3435­39 40­4445­49 50­5455­59 60­6465­69 70­74 75­7980­84 age (years) age (years) Pattern 7 0.5 0.5 1.5 ) x q( n 2.5 In 3.5 4.5 5.5 0 1­4 5­9 10­1415­1920­2425­29 30­3435­39 40­4445­49 50­5455­59 60­6465­69 70­74 75­7980­84 age (years) male female Source: INDEPTH Network 2002. Levels and Patterns of Mortality at INDEPTH Demographic Surveillance Systems | 83 increase that might be expected when HIV is having such an Without additional information, it is not possible to specu- easily observed impact on adult mortality. late on what may be producing this unique pattern. The male pattern is most similar to the Coale-Demeny North Pattern 4. Pattern 4 is a variation on pattern 1 with the model pattern, and the female pattern is closest to the important difference manifested in the 35 to 69 age range. Coale-Demeny West model, both of which embody high At all other ages, patterns 1 and 4 are negligibly different mortality in the same age ranges (INDEPTH Network except that infant and child mortality in pattern 4 is consis- 2002). They deviate from those patterns in that infant mor- tently slightly lower than pattern 1. But between ages 35 and, tality is substantially less than would be found in either roughly, 69 pattern 4 reveals significantly higher mortality model pattern, and child and adolescent mortality is signif- than pattern 1. This pattern is most similar to the UN gen- icantly higher: this might be called the "Super North" eral pattern for females and UN Latin American pattern for pattern. males (INDEPTH Network 2002). As was the case with pat- tern 1, most of the data contributing to pattern 4 comes Pattern 7. Pattern 7 is the other additional pattern identi- from West Africa. fied in the female data, and it, too, is interesting. It is derived from two sites located in central and West Africa. The reason Pattern 5. The HIV/AIDS pattern of mortality is most it was identified in the female data is obvious; there is a clearly visible in pattern 5. The data contributing to pattern substantial bulge in the female age pattern between ages 25 5 are derived from the three Tanzanian sites run by the Adult and 44. This pattern comes mainly from one small site in Morbidity and Mortality Project (AMMP) in Dar es Salaam, Zambia, so it is probably not worth extensive interpretation. Hai, and Morogoro (INDEPTH Network 2002). There is a The corresponding male pattern is similar to pattern 6, and striking bulge in the mortality of males between the ages of both are similar to the Coale-Demeny North model pattern 20 and 54 and for females between the ages of 15 and 49. (INDEPTH Network 2002). The North model pattern con- The female bulge is significantly narrower and more pro- tains relatively high child and teenage mortality coupled nounced, corresponding to the earlier infection of the with comparatively low mortality at older ages. This is con- female population and within a tighter age range than the sistent with the fact that malaria is an important contribu- male population. This pattern is not particularly similar to tor to mortality in both sites. any of the existing model patterns, but it is most closely matched with the UN General (female) and Latin American (male) model patterns (INDEPTH Network 2002). CONCLUSION Pattern 5 differs from pattern 3 mainly in the shape of the HIV/AIDS impact. The effect is more diffuse with age in In their 1996 review of demographic models, Coale and pattern 3, meaning that mortality is elevated through a Trussell (1996) describe the identification of "empirical broader age range, the magnitude of the elevation is more regularities" as one of the central motifs of demographic consistent, and the differences between the sexes are less modeling. Their term accurately describes the formulation apparent. Pattern 3 is derived largely from the Dar es Salaam of model mortality schedules and hints strongly at the lim- data, and this may reflect the fact that the epidemic is more its of their usefulness. Because there are no general, reliable, mature in Dar es Salaam and has consequently had enough theory-driven models of the individual risk of death, we time to infect a wider age range of people of both sexes. As must rely on commonly observed patterns to guide us-- with pattern 3, it is worth noting that infant and child mor- empirical regularities. Because these are simply commonly tality do not appear to be substantially affected in a manner observed patterns, however, we must be careful not to inap- comparable to adult mortality. propriately overinterpret them or substitute them for a thorough understanding of the mechanisms underlying the Pattern 6. Pattern 6 is one of the two additional patterns risks of death that lead to each pattern. Beyond being the identified in the female data. It is an interesting pattern that first step in understanding these mechanisms, identifying reveals high mortality of children and teenagers and com- underlying empirical regularities in age patterns of mortal- paratively low mortality of infants and adults of all ages. ity is still a useful exercise in its own right. This pattern is exhibited by sites in northeastern Africa and With fertility and to a lesser degree migration, mortality West Africa, with most of the data coming from Ethiopia. shapes the size and age composition of a population and is 84 | Osman A. Sankoh, Pierre Ngom, Samuel J. Clark, Don de Savigny, and Fred Binka consequently of great interest to policy makers who must no alternative. It is to begin addressing this need that the adequately plan for the future of a population. Mortality is INDEPTH Network has compiled a database of primarily one of the most valuable indicators of the health of a popu- African empirical mortality data and has begun to identify lation and, as such, is of critical importance to those who common age patterns of mortality described by those data manage health care infrastructures. As a sensitive indicator (INDEPTH Network 2002). of where a population is and where it will be in the near DSS site data are sometimes challenged as being unrepre- future, knowledge of mortality is a vital component of pop- sentative (Lopez et al. 2000) of surrounding constituencies ulation planning and management. due to the history of clinical and community-based inter- Existing systems of model life tables include systems vention trials that are conducted in such populations from developed by the Population Branch of the United Nations time to time. However, there is little evidence that the mor- Department of Social Affairs (1955) in the early 1950s, by tality patterns in trial sites are markedly different from those Coale and Demeny (1966) in the early 1960s, by Ledermann prevailing in similar settings nearby, where trials have not and Breas (1959) in the late 1950s, by Brass (1971) in the late been conducted. The reason for the paucity of evidence is 1960s, and finally by the UN again in the early 1980s (United that most trials are conducted on a subset of the population, Nations Department of International Economic and Social and they target a specific cause of mortality. Usually the Affairs 1982). Only two are in common use today; the Coale interventions do not continue after the trial, unless they and Demeny system and the 1982 UN set. However, neither are adopted as broad implementation policy (for example, of these systems was formulated with anything more than insecticide-treated nets), whereby the nontrial areas are, trivial data from Africa: apart from the West pattern that within a few years, at the same level of coverage for the inter- appears to contain a small amount of data from the white vention. We must also appreciate that our health system population of South Africa, none of the Coale and Demeny interventions have relatively weak modulating effects on patterns is based on data from Africa; and apart from a small population mortality compared with secular trends in amount of data from Tunisia, neither are the recent UN socioeconomic circumstances and environmental forces. patterns. We should recognize that the modest population size of a The lack of empirical data from Africa underlying either DSS site does not constitute a major flaw, however, as even of the commonly used model mortality systems is a signifi- sites monitoring small-sized populations can produce cant problem for demographers working in Africa, as robust measures of age-specific mortality when data are Preston, Heuveline, and Guillot (2001) eloquently state: "A aggregated over a period of several years. Moreover, data limitation, common to all the systems discussed so far collected over long periods of time from the same popula- [excluding the 1982 UN models], is that their empirical tion living in the same area can reveal important age- basis consists almost exclusively of the experience of devel- specific trends in the risk of death. Furthermore, when data oped countries. Most applications of model life tables, from a number of widely dispersed sites are brought however, are addressed to developing countries with incom- together, they provide both a geographically and temporally plete data." representative measure of mortality conditions. At present, Making this problem worse is the dissimilarity of the epi- only DSS sites provide data that can be used to depict the demiological profile of large parts of Africa to that of the temporal and geographical contours of mortality patterns developed world; it is clear that malaria and HIV are prima- in Africa. rily to blame. Both increase overall levels of mortality and may affect certain age groups disproportionately. The signif- icant endemic prevalence of these two diseases is responsi- NOTES ble for vast excesses in mortality in some of the harder hit age groups; malaria for young children and HIV for children 1. 1q0 represents the probability of dying before age one for newborns; 4 1 q represents the probability of dying before age five for those who sur- and young-to-middle-aged adults. vive to age one. In general, nqx is the probability of dying between the ages It is clear that the existing model mortality patterns are of x and x n for those who survive to age x. 2. q represents the probability of dying before age five for newborns. not appropriate for use in Africa, given that they were not 5 0 3. See INDEPTH Network 2002 for a discussion of why the age pattern built using data from Africa and that there is good reason to and level of mortality are orthogonal in the principal components model. This allows the clustering technique to identify clusters based on age believe that the mortality profile of Africa is different from pattern only, regardless of the level of mortality; essentially the principal other parts of the world. Until now, however, there has been components model controls for level. Levels and Patterns of Mortality at INDEPTH Demographic Surveillance Systems | 85 REFERENCES Assessment of Progress Toward Targets Based on Verbal Autopsy." Lancet 3: 349­58. Alonso, P. L., S. W. Lindsay, J. R. Armstrong, M. Conteh, A. G. Hill, P. H. Ledermann, S., and J. Breas. 1959. "Les dimension de la mortalité." David, G. Fegan et al. 1991. "The Effect of Insecticide-Treated Bed Nets Population 14 (4): 637­82. on Mortality of Gambian Children." Lancet 337: 1499­502. Lopez, A., J. H. Salomon, O. B. Ahmad, C. J. L. Murray, and D. Mafat. 2000. Armstrong Schellenberg, J. R., S. Abdulla, H. Minja, R. Nathan, O. Mukasa, "Life Tables for 191 Countries: Data, Methods, Results." GPE T. Marchant, H. Mponda et al. 1999. "KINET: A Social Marketing Discussion Paper 9, WHO, Geneva. Programme of Treated Nets and Net Treatment for Malaria Control in MacLeod, B., J. F. Phillips, and F. Binka. 1995. Analysis of Longitudinal Data Tanzania, with Evaluation of Child Health and Long-Term Survival." from the HRS. New York: Population Council. Transactions of the Royal Society of Tropical Medicine and Hygiene 93: Nevill C. G., E. S. Some, V. O. Mung'ala, W. Mutemi, L. New, K. Marsh, 225­31. C. Lengeler, and R. W. Snow. 1996. "Insecticide-Treated Bednets Binka, F. N., A. Kubaje, M. Adjuik, L. A. Williams, C. Lengeler, G. H. Reduce Mortality and Severe Morbidity from Malaria among Children Maude, G. E. Armah et al. 1996. "Impact of Permethrin Impregnated on the Kenyan Coast." Tropical Medicine and International Health 1 (2): Bednets on Child Mortality in Kassena-Nankana District, Ghana: A 139­280. Randomized Controlled Trial." Tropical Medicine and International Phillips, J. F., B. MacLeod, and B. Pence. 2000. "The Household Health 1 (2): 147­54. Registration System: Computer Software for Rapid Dissemination of Brass, W. 1971. "On the Scale of Morality." In Biological Aspects of Demographic Surveillance Systems." Demographic Research 2(6). Demography, ed. W. Brass. London: Taylor and Francis. Preston, S. H., P. Heuveline, and M. Guillot. 2001. "Modeling Age Patterns Brass, W., A. J. Coale, P. Demeny, D. F. Heisel, F. Lorimer, A. Romaniuk, of Vital Events." In Demography: Measuring and Modeling Population and E. van de Walle. 1973. The Demography of Tropical Africa. Processes, 191­210. Oxford: Blackwell. Princeton, NJ: Princeton University Press. Ross D. A., B. R. Kirkwood, F. N. Binka, P. Arthur, N. Dollimore, S. S. Breman, J. G., A. Egan, G. T. Keusch. 2002. "The Intolerable Burden of Morris, R. P. Shier, J. O. Gyapong, and P. G. Smith. 1995. "Child Malaria: A New Look at the Numbers." American Journal of Tropical Morbidity and Mortality Following Vitamin A Supplementation in Medicine and Hygiene 64 (1, 2): iv­vii. Ghana: Time Since Dosing, Number of Doses, and Time of Year." American Journal of Public Health 85 (9): 1246­51. Coale, A., and P. Demeny. 1966. Regional Model Life Tables and Stable Populations. Princeton, NJ: Princeton University Press. Tanzania Ministry of Health. 1997. "Policy Implications of Adult Morbidity and Mortality--End of Phase 1 Report." In Adult Morbidity Coale, A., and J. Trussell. 1996. "The Development and Use of and Mortality Project. Dar es Salaam: Government of Tanzania. Demographic Models." Population Studies 50 (3): 469­84. Timaeus, I. M. 1998. "Impact of the HIV Epidemic on Mortality in de Savigny, D., P. Setel, H. Kasale, D. Whiting, G. Reid, H. M. Kitange, Sub-Saharan Africa: Evidence from National Surveys and Censuses." C. Mbuya, L. Mgalula, H. Machibya, and P. Kilima. 1999. Linking AIDS 12 (Suppl. 1): S15­27. Demographic Surveillance and Health Service Needs--the AMMP/ United Nations Department of Social Affairs. 1955. Age and Sex Patterns of TEHIP Experience in Tanzania. Dar es Salaam. Mortality: Model Life Tables for Underdeveloped Countries. New York: Estève, J., E. Benhamou, and L. Raymond. 1994. Statistical Methods in United Nations. Cancer Research. Vol. 4, Descriptive Epidemiology. Lyons: IARC. United Nations Department of International Economic and Social Affairs. INDEPTH Network. 2002. Population, Health and Survival at INDEPTH 1982. Model Life Tables for Developing Countries. Population Studies Sites. Vol. 1 of Population and Health in Developing Countries. 77. New York: United Nations. Ottawa: International Development Research Centre. WHO/UNICEF (World Health Organization/United Nations Children's Korenromp, E. L., B. G. Williams, E. Gouws, C. Dye, and R. W. Snow. 2003. Fund). 2003. The Africa Malaria Report--2003. Geneva: WHO and "Measurement of Trends in Childhood Malaria Mortality in Africa: An UNICEF. 86 | Osman A. Sankoh, Pierre Ngom, Samuel J. Clark, Don de Savigny, and Fred Binka Chapter 8 Trends and Issues in Child Undernutrition Todd Benson and Meera Shekar Child undernutrition remains one of Africa's most funda- those currently achieved. Childhood and maternal under- mental challenges for improved human development. nutrition is the primary risk factor contributing to almost Because the time and capacities of caregivers are limited, far 30 percent of the estimated annual burden of disease on too many children on the continent are unable to access and the continent (table 8.1). The continuing human costs of effectively use at all times the food and health services they undernutrition for individual Africans as measured by its need for a healthy life. An estimated 200 million people on contribution to high mortality and morbidity and unreal- the continent, both children and adults, are undernour- ized human potential are enormous, and the aggregate costs ished, their numbers having increased by almost 20 percent at the national level impose a heavy burden on efforts to since the early 1990s (FAO 2003). This undernutrition starts foster sustained economic growth and improved human early in life--more than a third of African children under development. the age of five are stunted in their growth and must face a It is important to recognize the links between undernu- range of physical and cognitive challenges not faced by their trition, poverty, and, at the aggregate level, broad economic better-nourished peers. Malnutrition underlies 55 percent growth and national development. A necessary component of all deaths of children under five years of age globally for the success of new initiatives in Africa against poverty (Pelletier et al. 1994), and undernutrition is the major must be a sharp reduction in the current high levels of risk factor for over 28 percent of all deaths in Africa--some undernutrition. Broad-based economic growth is necessary 2.9 million deaths annually (Ezzati et al. 2003). Further, to increase incomes and consumption and thereby to reduce Pelletier and Frongillo (2003) have shown that had coun- poverty. Economic growth can primarily be achieved tries in Sub-Saharan Africa reduced undernutrition rates through enhanced economic productivity. Enhanced pro- over the period 1975­95 at the same modest rates as other ductivity comes about through broad improvements in the regions of the world, the region would have experienced intellectual, physical, and technical capacity of the popula- declines in under-five mortality rates 28 percent lower than tion. The potential intellectual, physical, and technical 87 Table 8.1 Percentage of Total Estimated Annual Burden cognitive impairment (Horton 1999). Iron deficiency in of Disease in Africa Attributed to Major Risk Factors adults has been estimated to decrease productivity between Childhood and maternal undernutrition 29.5 5 and 17 percent, depending on the nature of the work per- Underweight 18.0 formed (Horton 1999). Data from 10 developing countries Iron deficiency 2.9 have shown that the median loss in reduced work capacity Vitamin A deficiency 4.7 associated with anemia in adults is equivalent to 0.6 percent Zinc deficiency 3.9 of GDP, whereas an additional 3.4 percent of GDP is lost as a result of the effects on cognitive development attributable Other nutrition-related risks 3.0 to anemia in children (Horton and Ross 2003). The impact High blood pressure 1.3 of iodine deficiency diseases on cognitive development High cholesterol 0.6 alone has been associated with productivity losses totaling High body-mass index (BMI) 0.4 approximately 10 percent of GDP (Horton 1999). Low fruit and vegetable intake 0.4 It is only when Africans have improved nutrition security Physical inactivity 0.3 that we will begin to see sustained progress toward the Sexual and reproductive health risks 20.1 achievement of the aspirations laid out in the Millennium Unsafe sex 19.4 Development Goals (MDG), in the plans under the New Lack of contraception 0.8 Partnership for Africa's Development (NEPAD), and in the Environmental risks 10.1 many national poverty reduction strategies. Unsafe water, sanitation, and hygiene 5.3 Indoor smoke from solid fuels 3.5 Other 1.1 WHAT CAUSES UNDERNUTRITION? Occupational risks 0.7 The United Nations Children's Fund (UNICEF) conceptual Addictive substances 2.9 framework of the determinants of the nutritional status of Source: Ezzati et al. 2003. children shown in figure 8.1 presents a generalized under- Note: Percentage of total estimated annual burden of disease is the share of disability-adjusted life years (DALYs) lost attributed to major risk factors. Total estimated DALYs lost annually in standing of how undernutrition is the outcome of specific Africa is 349,513,000. development problems related directly to the dietary intake and the health status of the individual. The quality of these capacity of the population is dependent on improved nutri- immediate determinants, in turn, is determined by the tion, particularly for young children and women in their underlying food security status of the household in which a childbearing years. Similarly, the effective use of such capac- child resides. However, of equal importance is the availability ity is dependent on a properly nourished population in of health services and a healthy environment and the quality which individuals are living healthy and active lives and are of care the child receives--that is, whether the available able to creatively contribute to their own and the nation's dietary resources for good nutrition are used effec- economic well-being. tively through appropriate caring practices. Sustained Malnutrition in any stage of childhood affects schooling healthy and active life is only possible when these underly- and, thus, the lifetime earnings potential of the child. ing determinants--food, health, and care--are each maxi- Malnutrition affects educational outcomes, including a mized. None of these is sufficient in itself, but all are neces- reduced capacity to learn (as a result of early cognitive sary for good child growth. deficits or lowered current attention spans) and fewer total The degree to which these three underlying determinants years of schooling. In Zimbabwe, stunting, via its association are expressed positively or negatively is a question of with a 7-month delay in school completion and a 0.7-year resources. These resources include the availability of food loss in grade attainment, has been shown to reduce lifetime but extend much further to include the physical and eco- income by 7 to 12 percent (Alderman, Hoddinott, and nomic access that a child or his or her caregiver has to that Kinsey 2003). In general, in low-income agricultural coun- food, the caregiver's knowledge of how to use available food tries, the physical impairment associated with malnutrition and to properly care for the child, the caregiver's own health is estimated to cost 2 to 3 percent of gross domestic product status, and the control the caregiver has over resources with- (GDP) per year, even without considering the long-term in the household that might be used to nourish the child. productivity losses associated with developmental and Finally, the level of access to information on and services for 88 | Todd Benson and Meera Shekar Figure 8.1 The UNICEF Conceptual Framework of the Determinants of Nutritional Status Outcome Nutritional status Immediate Dietary Health determinants intake status Healthy Household Quality of environment, food security care health services Underlying determinants Poverty Food security Caregiver Resources for constrains the resources resources health level of these determinants for · quantity food · knowledge and · availability of individual produced access to public health households. · quality food education services produced, diet · health status · sanitation, diversity · control of access to · cash income resources clean water · food transfers Basic Institutions determinants The impact that the resources potentially Political and ideological framework available to the household have on nutritional status is mediated and Economic structure constrained by overarching economic, political, and institutional Potential resources structures. Human, agro-ecological, technological Sources: Jonsson 1993; Smith and Haddad 2000; and UNICEF 1990. maintaining health, whether preventive and curative health issue of immediate concern to any national development services are available, and the presence or absence of a strategies. healthy environment with clean water, adequate sanitation, Finally, it is important to recognize from this framework and proper shelter all contribute equally to determining the that a household or nation being food secure is not in itself nutritional status of a child. sufficient to ensure the good nutritional status of children When the distribution of resources within society is cen- and others within that household or nation. It is possible to tral to accounting for why some are undernourished and have poor nutritional outcomes without being food insecure. others are not, the framework moves from the realm of the Food security is concerned with physical and economic access individual and household to the political. The framework to food of sufficient quality and quantity. However, one may links the availability of nutrition resources to a set of basic have reliable access to the components of a healthy diet, but determinants, which are themselves a function of how soci- because of poor health or care (such as poor infant-feeding ety is organized in regard to economic structure, political practices), ignorance, or personal preferences, one may not and ideological expectations, and the institutions through be able or may choose not to use the nutrients to which one which activities within society are regulated, social values has access. In parallel to food security, one can speak of nutri- are met, and potential resources are converted into actual tion security. Nutrition security is achieved for a household resources. Consequently, this conceptual framework identi- when secure access to food is coupled with a sanitary envi- fies undernutrition as a subject for political debate and an ronment, adequate health services, and knowledgeable care. Trends and Issues in Child Undernutrition | 89 INDICATORS OF CHILD UNDERNUTRITION: growth are identified by comparison with the growth OUTCOMES AND INPUTS characteristics of children of a similar age, disaggregated by sex, in a standard, nutritionally secure population. In this section, measures of child undernutrition and of its Children whose growth is more than two standard devia- determinants--the outcomes and inputs, respectively, to tions (Z-scores) below the mean physical characteristics for the process determining nutritional status--are presented the nutritionally secure reference population are considered as a foundation for the discussion on the current status and undernourished.1 trends in these measures in Sub-Saharan Africa. Of the other indicators of aggregate nutritional status, infant and under-five mortality and low birthweight preva- lence are important proxy measures of nutrition security. Outcome Indicators Low birthweight reflects fetal growth retardation due to the Table 8.2 provides a summary of the principal measures poor health and nutrition of the mother and also serves as of both individual and aggregate nutritional status. Of an indicator of risk of infant mortality and future poor these, the most commonly employed are the anthropo- health. From a life cycle perspective on nutrition and the metric measures of child nutritional status--stunting, intergenerational transmission of poverty within society, as underweight, and wasting. Children with abnormally low presented in figure 8.2, low birthweight is a critical measure. Table 8.2 Indicators of Nutritional Status Indicator How collected Comments Low height-for-age, "stunting" Height (or length) measurements of children 6 to Indicative of long-term nutritional status of children. Best measure of 60 months or 6 to 36 months in age. Children with cumulative growth retardation. abnormally low growth are identified by comparison to physical characteristics of children of similar age, disaggregated by sex, in a standard, nutritionally secure population. Low weight-for-age, Weight measurements of children 6 to 60 months or Nonspecific indicator of overall undernutrition--measures a "underweight" 6 to 36 months in age. Analysis similar to that above. combination of chronic and acute undernutrition. Low weight-for-height, Weight and height measurements of children 6 to Measures acute child undernutrition. Indicative of sharp short-term "wasting" 60 months or 6 to 36 months in age. Analysis similar fluctuations in nutritional status. Most useful in emergencies where to that above. severity of the nutritional crisis is being assessed or short-term progress in nutritional status is being monitored. Mid-upper arm circumference Distance around the mid-upper arm. Standard threshold Used in emergencies in contexts similar to where the wasting values used to determine whether a child is at risk. measure would be used. Body mass index Computed as weight (in kg) divided by height Indicator of nutritional status in adults and older children. Only (in meters) squared. Standard threshold values used to indicator commonly used to assess both short- and long-term determine whether an individual is underweight or nutritional status of adults, usually women of childbearing age. Used overweight. for assessing the prevalence of both underweight and overweight individuals in a population. Micronutrient deficiency Based on biochemical analysis (laboratory, some field Difficult to collect on a large scale. Biochemical measures require tests) or symptomatic diagnosis of deficiency-related access to analytical laboratory facilities. For some micronutrients, disease--night blindness (vitamin A), goiter (iodine), most notably zinc, biochemical measures of nutrient deficiency are not anemia (iron). available and clinical diagnosis is not possible. Infant mortality (under one Census, Demographic and Health Surveys. Proxy measures of aggregate nutritional status. Integrated measure of year) and under-five mortality risks to child survival, among which poor nutritional status is a major rates factor. Low birthweight prevalence Proportion of babies born weighing less than 2,500 g. Low birthweight reflects fetal growth retardation due to the poor Collected from health statistics. health and nutrition of the mother. However, also serves as an indicator of risk of infant mortality and future poor health. Predicts, less precisely, economic and broad human potential of the child. From a life cycle perspective on nutrition and the intergenerational transmission of poverty within society (see figure 8.2), low birthweight is a critical measure. Source: Compiled by authors. 90 | Todd Benson and Meera Shekar Figure 8.2 The Burden of Undernutrition through the Life Cycle and across Generations Higher Impaired mental Reduced capacity to mortality rate development care for child Increased risk of adult chronic disease BABY OLDER PEOPLE Untimely/inadequate feeding Low birthweight Malnourished Frequent infections Inadequate food, Inadequate health, and care catch-up growth Inadequate Inadequate food, infant Inadequate fetal health, and care nutrition nutrition CHILD Stunted Reduced Reduced physical WOMAN mental labor capacity, Malnourished capacity lower educational attainment, restricted economic PREGNANCY Inadequate food, potential, Low weight gain health, and care shortened life ADOLESCENT expectancy Stunted Higher maternal Inadequate food, mortality health, and care Reduced physical labor capacity, lower educational attainment Source: Adapted from SCN 2000. Finally, measures of micronutrient deficiency are critical Among the most common measures of access to food at for assessing undernutrition in Africa. The burden of disease the national level is the Food and Agriculture Organization of attributed to micronutrient deficiencies is often referred to the United Nations (FAO) undernourishment measure, as "hidden hunger": The clear link between a lack of suffi- which takes into account food availability from production cient food to eat and a poor physical state that one sees with and trade and, using the distribution of consumption levels a lack of carbohydrate, protein, and fat in the diet is not as across the population, estimates what proportion of the readily seen when considering micronutrient deficiency. population is unable to meet its daily energy requirements. Subclinical levels of deficiency can have serious and Similar assessments can be made using detailed house- irreparable consequences on health, mortality, and eco- hold consumption surveys. Dietary quality assessments can nomic productivity. There are four principal micronutrient be made using information from both commodity- deficiencies of public health concern in Africa--vitamin A, disaggregated national food balance sheets and household iron, zinc, and iodine. Table 8.3 shows their principal surveys. Although these estimates are by nature imprecise, sources and the health effects of deficiency. and there are many valid critiques of them, in the absence of better indicators they are used extensively to monitor food security status. Input Indicators Direct measurement of quality of care is more problem- The three underlying determinants of child undernutrition atic, requiring the use of proxy measures. Female education are household food security; quality of care; resources for and literacy levels are important elements in this regard. health, including access to health services; and a healthy There is considerable evidence that the nutritional status of environment (figure 8.1). These three inputs to the process children varies directly with the level of education of their determine nutritional outcomes both for individuals and parents, and in particular, their mothers. Women who have for populations as a whole. It is important to recognize that more education are better able to understand important balance is required across these factors if undernutrition is nutritional care practices and to follow them with the chil- to be sustainably reduced. dren in their care. An additional contribution to nutritional Food availability, access to food, and diet quality are care is the availability of public health services, particularly various dimensions of food security that can be measured. for prenatal care and infant health and growth monitoring. Trends and Issues in Child Undernutrition | 91 Table 8.3 Effects of Deficiency of Micronutrients and Principal Dietary Sources Micronutrient Principal dietary sourcesa Health effects of deficiency Vitamin A Breast milk, liver, egg yolk, milk and dairy products, green leafy In deficient populations, improvements in vitamin A status are vegetables (esp. kale, amaranth, sweet potato, cowpea, and associated with a 23 percent overall reduction in mortality among cassava leaves), yellow- and orange-colored fruits and vegetables under-five children. Night blindness, the first stage in a set of (carrots, pumpkin, mango, papaya, oranges), orange-fleshed sweet increasingly severe eye problems (xerophthalmia) that leads to potato, red palm oil. corneal ulcers and to blindness. Impaired resistance to infection. Iron Liver, meat, poultry, fish, cereals (esp. whole grain), nuts, beans, Anemia, especially in women and children. Fatigue, with adverse and green leafy vegetables. Also commercially produced iron- effects on learning, productivity, and earnings. Pregnancy complica- fortified foods. tions, maternal mortality, premature birth, and low birthweight. In children, significant loss of cognitive abilities as well as decreased physical activity and reduced resistance to disease. Significant impact on school enrollment and adult productivity. Zinc Animal and fish products, beans, and other legumes. Low birthweight and poor growth, reduced resistance to infectious diseases, increased incidence of stillbirths, and possibly impaired cognitive development. Iodine Underlying cause of iodine deficiency is a deficiency of iodine in Mental retardation and stunted growth among children-- the local soil on which vegetation grows, animals graze, and crops "cretinism." Goiter is a symptom of iodine deficiency, which in are cultivated. This results in a shortage of iodine in local food- itself may pose social, economic, and physiological burdens on the stuffs. Iodine fortification is the principal source of iodine, usually individual. Improvements in iodine status are linked to a 13-point through iodized salt. improvement in IQ among children. Iodine deficiency is linked to reduced economic productivity. Source: Compiled by authors. a. Humans generally are less able to absorb and utilize the vitamin A, iron, and zinc that come from plant foods than those that come from animal and fish sources. Although the availability of such services is an important determinants considered in the previous paragraphs--for measure in itself, measures of the degree to which caregivers example, increased income will increase access to food, to adopt proper nutritional care practices--breastfeeding, health care, and to child care resources--and emerge, in other child-feeding practices, and health-seeking behavior, part, from the basic determinants beneath the underlying in particular--can also be generated using Demographic determinants of the framework. Moreover, as was noted ear- and Health Surveys (DHS). Finally, proxy measures for the lier, improved nutrition is itself an important determinant relative control that women are able to exercise over their of economic growth and improved welfare--the relation- own and their household's income and other resources have ship between these welfare measures and child nutritional also been shown analytically to be significantly correlated status is neither linear nor one way. Nevertheless, they are with the nutritional status of children in the household important, relatively nonspecific inputs into the nutritional (Smith et al. 2003). The more control women have over such process. resources, the more likely it is that those resources will be used to provide proper nutritional care to children in the Data Sources household. A broad range of measures can be used to assess the The principal data sources for nutritional outcome indica- degree to which good nutrition is assured by available tors for Sub-Saharan Africa come from DHS surveys carried health services and a healthy environment. These include out by national statistical offices. These surveys also contain access to health facilities and health professionals, the bur- considerable information on nutritional caring practices. den of disease that health systems must bear, immunization The World Health Organization (WHO) has compiled child coverage, the percentage of mothers receiving prenatal care, undernutrition data from these and other surveys into a the use of safe drinking water and sanitation facilities, and global database on child growth and malnutrition (WHO so on. 2003). Other data sources for nutrition-related indicators Finally, general and household welfare measures-- include national population and agricultural censuses for income, assets, and poverty--are also important in account- demographic and agricultural production information, ing for the nutritional status of children. Within the household economic surveys for welfare and consumption conceptual framework, these cut across the underlying measures, and facilities surveys for access to social services. 92 | Todd Benson and Meera Shekar STATUS AND TRENDS IN CHILD Figure 8.3 Prevalence of Stunting in Children Age 6 to UNDERNUTRITION 60 Months, by Country (percent) Although poor nutrition and hunger affects all ages, the 30° 20° 10° 0° 10° 20° 30° 40° 50° IBRD 34519 MARCH 2006 long-term development and welfare implications are espe- Mediterranean Sea cially important for young children because most undernu- 30° 30° trition happens in the womb or in the first two years of life. Much of this early damage--both physical growth and brain Red 20° Sea 20° development--is irreversible (SCN 2004, p. 40). Moreover, as the nutritional status of these children when in the womb 10° 10° is an important determinant of their developmental poten- tial, maternal undernutrition is of equal concern. Figure 8.2 shows how undernutrition can be perpetuated across gener- 0° ATLANTIC 0° OCEAN INDIAN ations from mother to child in a spiral of poverty and OCEAN despair. Consequently, sustainably enhancing the nutri- 10° 10° tional status of young children and their mothers is the key intervention in breaking this cycle to enable each succeeding 20° 20° less than 25 generation to aspire to an increasingly healthy life. 25 ­ 40 more than 40 In most of Sub-Saharan Africa increased levels of under- no data 30° 30° international boundaries nutrition are strongly associated with the age of the child. The decline in nutritional status from birth is astonishingly This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. swift as newborns in many African households face a 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° challenging health environment and, in many cases, receive Source: Data from UNICEF 2003. suboptimal feeding, that is, lack of exclusive breastfeeding and inappropriate or untimely complementary feeding. in figure 8.3. In contrast to North Africa, where child Although the average nutritional status of newborns in undernutrition appears to be addressed relatively effectively, these countries is similar to that of the well-nourished refer- in Sub-Saharan Africa the pattern is less encouraging and ence population, by about the age of 12 to 15 months about somewhat more complex. Coastal West Africa, central half of the children are underweight (weight-for-age Z-score Africa, and southern Africa have lower rates of stunting. less than 2.0) with very little improvement thereafter. This Landlocked countries and others with a large proportion of growth retardation experienced in the first year of life is very their population in the interior tend to have higher rates. difficult to overcome in the later years of childhood. General perceptions that developed in the 1980s of poorly Globally, progress is being made in reducing undernutri- nourished populations in the Sahel and Ethiopia remain rel- tion. The prevalence of child undernutrition has declined atively accurate today. The grouping of countries with the significantly over the past 25 years. Rates of stunting (low highest prevalence of stunting, however, is found in south- height-for-age) among children age six months to five years ern and eastern Africa, reflecting a complex set of challenges in all developing countries dropped almost 20 percentage that include civil conflict, economic downturns due to points from 49 to 30 percent between 1980 and 2000, where- macroeconomic mismanagement or commodity price as underweight (low weight for age) rates dropped from 38 shocks, and droughts and floods, or the legacies of such to 25 percent (de Onis et al. 2004). Taken as a whole, how- events. ever, Africa is an unfortunate exception to these trends. Over The prevalence of undernutrition can also be assessed at the period 1980 to 2000, stunting rates declined by less than a subnational level. Using data from a range of DHS surveys, 4 percentage points in Africa, so that, with population the Center for International Earth Science Information growth, the actual number of stunted children actually Network (CIESIN) developed a series of maps describing increased by more than 12 million to 48.5 million. Both the prevalence of undernutrition at this scale across the con- relative and absolute numbers of underweight children in tinent. Two of these are shown in figure 8.4. The map on the Africa increased over the same period. left of the figure is comparable in form to the map shown in National and broader regional patterns can be seen in the figure 8.3 but shows underweight prevalence, instead of map portraying the national prevalence of stunted children stunting, and subnational spatial units. The map on the Trends and Issues in Child Undernutrition | 93 Figure 8.4 Subnational Estimates of the Prevalence of Underweight Preschool Children and Area Density of Underweight Children a. Percentage of underweight children b. Population density of underweight children per square kilometer 0° 30° 0° 30° IBRD 34520 MARCH 2006 Mediterranean Mediterranean Sea Sea 30° 30° 30° 30° Red Re d Sea Sea ATLANTIC OCEAN ATLANTIC 0° OCEAN 0° 0° 0° INDIAN INDIAN OCEAN OCEAN less than 10 less than 0.5 10 ­ 20 0.5 ­ 1 20 ­ 30 1 ­ 5 30 ­ 40 5 ­ 10 40 ­ 50 10 ­ 25 more than 50 more than 25 no data no data This map was produced by the Map Design Unit of The World Bank. The 30° 30° 30° boundaries, colors, denominations 30° international boundaries international boundaries and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of 0° 30° 0° 30° such boundaries. Source: Data from CIESIN 2004. right provides a different perspective on undernutrition. By important, safe water and adequate sanitation, health taking into account the area density of the population from services, and the information needed by mothers or other which the estimates of underweight are being made, the caregivers to provide children with effective care are relative- map provides an indication of the spatial intensity of the ly less accessible in rural areas. It is not correct, however, to problem of underweight children--"hotspots" of undernu- argue that undernourished children are primarily a rural trition. In this light, although problematic across much of phenomenon. In most of the countries examined in fig- subtropical Africa, the map shows undernutrition hotspots ure 8.5, at least one out of five urban preschoolers is stunted. in a handful of areas, including coastal and northern Child undernutrition in Africa can also be examined Nigeria, upland Ethiopia, the Lake Victoria basin, southern from the dimensions of the gender and the age of the child. Malawi, and central Mozambique. It is important to note In contrast to South Asia, for example, differences in the that in many of these hunger hotspots, food production is prevalence of child undernutrition according to gender are not the limiting factor. One such example is the Iringa not strong. In Sub-Saharan Africa, the tendency is for boys region in Tanzania, where over 70 percent of the children are to have a somewhat higher probability of being undernour- stunted in their growth, even though it is a food basket for ished than girls (Svedberg 1990). For example, in an assess- Tanzania. ment of stunting rates for children age 6 to 36 months in Place of residence--rural or urban--can yield another 28 African countries during the period 1987 to 2002, the important evaluation of the spatial pattern of child under- average difference in stunting rates between boys and girls nutrition in Africa. Figure 8.5 shows stunting prevalence was 2.6 percentage points, with stunting rates being higher for children under five, disaggregated by urban or rural among boys in all but four of these countries. residence. In all 17 countries, undernourished children are Yet another interesting dimension of undernutrition in more prevalent in rural areas than in urban centers. Sub-Saharan Africa is reflected in figure 8.6, which shows Although food is produced on farms in rural areas, this does that although undernutrition rates in these countries are not mean that rural children are better nourished. Equally higher among the poorer households, rates remain relatively 94 | Todd Benson and Meera Shekar Figure 8.5 Stunting Prevalence among Preschoolers, by Urban or Rural Residence, Selected Countries 60 50 40 30 children % 20 10 0 nin ad bon ana Be rkina Faso rundi Ch inea Mali anda anda Bu e d'Ivoire Ethiopia Ga Gh Gu Malawi Rw Tanzania Ug Zambia Zimbabwe Bu Côt Mauritania country urban rural Source: Data from ORC/Macro 2004, DHS survey. Figure 8.6 Prevalence of Underweight Preschool Children doses of vitamin A and iron for women and children. The by Wealth Quintile, Selected Countries simplest and most sustainable way to eliminate micronu- 50 trient deficiencies is to make sure that individuals and households know the importance of a balanced diet and 40 have access to what is required to consume such diets. However, in spite of the progress made in combating this prevalence 30 "hidden hunger," many Africans still consume insufficient 20 amounts of the relatively small quantities of these nutrients that they require or, due to poor health, are unable to use underweight % 10 effectively that which they do consume and continue to suf- fer from micronutrient deficiencies. High levels of anemia 0 Benin Ghana Malawi Mali Tanzania Uganda Zambia result in serious cognitive and productivity losses, reducing country the ability of women to work and provide adequate care poorest 2nd quintile 3rd quintile 4th quintile wealthiest for their children and making pregnancy and childbirth much more risky for them than would otherwise be the case. Source: Data from Gwatkin et al. 2003. Between 15,000 and 20,000 African women die each year owing to severe iron deficiency anemia. The high prevalence high among the nonpoor. In many countries in the region, of goiter in school-age children points to hundreds of thou- more than 15 percent of the children in the highest wealth sands of children in Africa who have lowered intellectual quintile are underweight. Given the significant proportion capacity as a result of iodine deficiency. Vitamin A deficien- of nonpoor children that are undernourished, we should cies in children are common across the continent, reducing recognize that improved household welfare in itself is not their ability to resist infection and contributing to the deaths sufficient to eliminate undernutrition. The higher level of of more than half a million African children annually nutritional resources available to wealthier households must (UNICEF and MI 2004). be used effectively through proper care if undernutrition is Child undernutrition is closely correlated with under-five to be eliminated in such households. mortality. For the countries south of the Sahara, most recent Progress has been made in Africa over the past 15 years estimates of under-five mortality are about 170 deaths in the in addressing micronutrient deficiency diseases. Although first five years of life for every 1,000 live births (UNICEF these problems are serious, clinical solutions are relatively 2003). Figure 8.7b shows that, although many countries inexpensive to implement--salt iodization, fortification of are making progress in reducing child deaths, some coun- commonly consumed commercial foods, and supplemental tries are failing to maintain effective past efforts and are Trends and Issues in Child Undernutrition | 95 Figure 8.7 Under-Five Mortality Rates and Progress Being Made in Reducing Under-Five Mortality, by Country a. Under-five mortality rate, 2002 b. Progress in reducing under-five mortality, 1960­2002 30° 20° 10° 0° 10° 20° 30° 40° 50° 30° 20° 10° 0° 10° 20° 30° 40° IBRD 34521 MARCH 2006 Mediterranean Mediterranean Sea Sea 30° 30° 30° 30° Red Re d Se 20° 20° 20° 20° a Sea 10° 10° 10° 10° 0° ATLANTIC 0° 0° ATLANTIC 0° OCEAN INDIAN OCEAN INDIAN OCEAN OCEAN 10° 10° 10° 10° accelerating decline 20° less than 100 20° 20° 20° 100 ­ 175 decelerating decline more than 175 stagnant or rising rate 30° no data 30° 30° no data 30° international boundaries international boundaries This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° Source: Data from UNICEF 2003. Note: a. Under-five mortality is the number of deaths in the first five years of life per 1,000 live births. b. Progress in reducing under-five mortality is shown as a comparison of the annual rate of reduction in under-five mortality during 1990­2002 with that during 1960­90. experiencing a decline in the rate of progress against under- Figure 8.8 Scatter Plot of National Under-Five Mortality and five mortality. In spite of the high levels of under-five mor- Underweight Prevalence Rates tality in Africa, only a handful of African countries show 50 accelerating declines in levels of under-five mortality. The scatter plot in figure 8.8 demonstrates the intertwining of (%) 40 nutritional and health concerns and highlights the scope of 30 the linked problems of undernutrition and consequent prevalence under-five mortality in Sub-Saharan Africa. If one uses an 20 underweight prevalence of 20 percent and an under-five mortality rate of 50 per 1,000 live births (bold gridlines in underweight 10 figure 8.8) as thresholds above which these issues should be seen as critical public health problems requiring urgent 0 50 100 150 200 250 300 action, under-five mortality remains a pressing issue in all under-five mortality (per 1,000 live births) countries, and in more than two-thirds, undernutrition is a Source: Data from UNICEF 2003. critical concern. Note: Countries with underweight prevalence below 20 percent: Botswana, Rep. of Congo, Djibouti, Equatorial Guinea, Gabon, Gambia, Lesotho, São Tomé and Principe, South As noted, approximately 55 percent of all child deaths in Africa, Sudan, Swaziland, Zimbabwe. developing countries can be attributed in part to undernu- Countries with underweight prevalence above 20 percent (serious public health concern): Angola, Benin, Burkina Faso, Burundi, Cameroon, Central African Republic, Chad, trition through its exacerbation of the effects of disease on a Comoros, Dem. Rep. of Congo, Côte d'Ivoire, Eritrea, Ethiopia, Ghana, Guinea, Guinea-Bissau, Kenya, Liberia, Madagascar, Malawi, Mali, Mauritania, Mozambique, Namibia, Niger, Nigeria, child's health (Pelletier et al. 1994). For example, 61 percent Rwanda, Senegal, Sierra Leone, Somalia, Tanzania, Togo, Uganda, Zambia. of diarrhea deaths and 57 percent of malaria deaths would not occur if it were not for the underlying undernutrition. underweight children (Black, Morris, and Bryce 2003). Even mildly underweight children have nearly double the Further, as noted earlier, Pelletier and Frongillo (2003) have risk of death of their well-nourished counterparts. This shown that had the countries of Sub-Saharan Africa reduced risk increases to five- to eightfold in moderately to severely undernutrition rates during 1975 to 1995 at the same 96 | Todd Benson and Meera Shekar Figure 8.9 Daily Dietary Energy Supply Available and Food Production Index, by Country a. Dietary energy supply, 1999­2001 (calories) b. Food production index, 1998­2000 30° 20° 10° 0° 10° 20° 30° 40° 50° 30° 20° 10° 0° 10° 20° 30° 40° IBRD 34522 MARCH 2006 Mediterranean Mediterranean Sea Sea 30° 30° 30° 30° Red Red Sea Sea 20° 20° 20° 20° 10° 10° 10° 10° 0° ATLANTIC 0° 0° ATLANTIC 0° OCEAN INDIAN OCEAN INDIAN OCEAN OCEAN 10° 10° 10° 10° 20° 20° 20° 20° more than 2,400 more than 140 2,100 ­ 2,400 110 ­ 140 less than 2,100 less than 110 30° 30° 30° 30° no data no data This map was produced by the Map Design Unit of The World Bank. international boundaries international boundaries The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° Source: Dietary energy data from FAO 2003; food production data from World Bank 2003. Note: a. Dietary energy supply (kcals/capita/day) was computed from total food production and food imports and exports. b. 100 aggregate value of national food crop production in 1989­91. modest rates as other regions of the world, the region would The two maps in figure 8.9 provide some indication of have experienced under-five mortality rates 28 percent current availability of food in Africa based on total dietary lower than those achieved. supply (from food balance sheet data) and an index of food production. The trends are not encouraging. A moribund food crop sector is found in combination with insufficient What Accounts for the Patterns food in quite a few countries, particularly in eastern and of Undernutrition Observed? central Africa. In such countries, it is likely that many rural Two-thirds of Africans reside in rural areas, with the major- households are unable to produce as much food as they ity engaged in crop and livestock production for both their require from their own land and cannot afford to go to the own use and market sale. Higher production on their own market to make up any deficiencies. Chronic food insecurity farm or from their own herds enhances the food security for results. Any future negative shock to a household's well- such households. For food purchasers, higher production being likely will propel it into deepening undernutrition and generally means lower food prices and access to a greater ill health. quantity of food in the market for a given income level. Since most undernutrition happens in the first two years Elsewhere around the globe other underlying determinants of age, when food security is less of an issue than are caring of nutritional status, such as women's education and its practices, such as exclusive breastfeeding and appropriate impact on the quality of care provided or the availability of complementary feeding (and whether women have enough health resources, are of greater importance, but in Africa time within their household and other economic duties to the relatively low food availability has been shown to be provide such care), caring practices play a key role in under- among the most important underlying determinants of nutrition in Sub-Saharan Africa. Moreover, the high under- the high levels of child undernutrition (Smith and nutrition rates found among the nonpoor suggest that even Haddad 2000, p. 84). Nevertheless, as mentioned earlier, when access to food is not constrained and caring practices many regions of Africa that are notable for their relatively might be expected to be relatively good, other factors, such high level of food production also suffer from high levels of as an unhygienic environment or poor health services, also undernutrition. may underlie poor nutrition outcomes. Trends and Issues in Child Undernutrition | 97 Figure 8.10 Access to Safe Water and Adequate Sanitation, by Country a. Percentage of rural population with access to b. Percentage of population with adequate sanitation, 2000 safe drinking water, 2000 30° 20° 10° 0° 10° 20° 30° 40° 50° 30° 20° 10° 0° 10° 20° 30° 40° IBRD 34523 MARCH 2006 Mediterranean Mediterranean Sea Sea 30° 30° 30° 30° Red Re d Se Sea 20° 20° 20° 20° a 10° 10° 10° 10° 0° ATLANTIC 0° 0° ATLANTIC 0° OCEAN INDIAN OCEAN INDIAN OCEAN OCEAN 10° 10° 10° 10° 20° 20° 20° 20° more than 70 more than 70 45 ­ 70 45 ­ 70 less than 45 less than 45 30° 30° 30° 30° no data no data This map was produced by the Map Design Unit of The World Bank. international boundaries international boundaries The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° Source: UNDP 2003. Figure 8.10 shows current progress in Africa in improv- pattern of wealth across the continent. Access to doctors and ing public health by providing clean water and adequate secondary and tertiary medical care facilities is as much a sanitation. Although several countries have made admirable question of resources as it is of medical need. Poorer coun- strides in ensuring that their citizens live in a hygienic envi- tries tend to have fewer doctors per capita. The map of ronment, most countries must continue their efforts. By prenatal care, however, is arguably the more useful for our comparing these maps with those in figure 8.9, one can purposes here. Prenatal care services are mainly provided identify several countries--Benin and Namibia, among as a component of primary health care programs. These others--where one should find reasonable access to food programs are also the a principal means of providing direct but where improved water and sanitation is clearly needed nutrition services to the population. Indeed, an important if sustainable reductions in child undernutrition are to be component of prenatal care is proper nutrition to prevent achieved. low birthweight. Overall, the picture across Africa concern- Closely linked to a hygienic environment is the health ing the provision of these health care services is encourag- status of the population more generally. Individuals suffer- ing. Still, although there is room for expansion--large num- ing from illness are unable to properly use the nutrients in bers of pregnant women still are receiving no attention-- the food to which they have access. Consequently, lack of and likely the quality of the care provided can be enhanced, access to health services--both preventive and curative--is a foundation upon which to build is in place in many a central cause of undernutrition. A broad range of indica- countries. tors can be used to assess this factor in Africa. In figure 8.11 Over the past two decades, HIV infection has added an two indicators of access to health care--number of doctors almost overwhelming additional burden to Africa's health per 100,000 people and percentage of pregnant women care services. As shown in figure 8.11c, countries in south- receiving prenatal health care--are presented, together with ern and eastern Africa are currently facing the most severe a map showing the prevalence of human immunodeficiency levels of infection, but all countries must put in place mech- virus (HIV) infection, an important challenge to African anisms to effectively control the disease and care for those health care systems. Figure 8.11a, which shows the number infected. Existing health services are and will be extremely of medical doctors per 100,000 persons, largely reflects the stretched, and without significant additional resources, the 98 | Todd Benson and Meera Shekar Figure 8.11 Health Services, by Country a. Number of medical doctors per 100,000 people b. Percentage of mothers receiving prenatal care 30° 20° 10° 0° 10° 20° 30° 40° 50° 30° 20° 10° 0° 10° 20° 30° 40° IBRD 34524 Mediterranean Mediterranean MARCH 2006 Sea Sea 30° 30° 30° 30° Red Red 20° Sea 20° 20° Sea 20° 10° 10° 10° 10° ATLANTIC ATLANTIC OCEAN 0° OCEAN 0° 0° 0° INDIAN INDIAN OCEAN OCEAN 10° 10° 10° 10° more than 75 more than 30 20° 50 ­ 75 10 ­ 30 20° 20° 20° less than 50 less than 10 no data no data 30° 30° 30° 30° international boundaries international boundaries 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° c. Prevalence of HIV-infected adults, age 15­49 years, end-2001 (percent) 30° 20° 10° 0° 10° 20° 30° 40° 50° Mediterranean Sea 30° 30° Re d 20° Sea 20° 10° 10° ATLANTIC 0° OCEAN 0° INDIAN OCEAN 10° 10° less than 5 This map was produced 20° 5 ­ 15 20° by the Map Design Unit of The World Bank. The more than 15 boundaries, colors, denominations and any other information shown no data on this map do not imply, 30° 30° on the part of The World international boundaries Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° boundaries. Sources: Data on medical doctors from UNDP 2003; data on prenatal care and HIV-infected adults from UNICEF 2003. overall quality of service provision, including nutrition- production both now and in the future, restricts the access oriented services, will suffer. Moreover, HIV has exacerbat- to available food and necessary health care that the house- ing effects on undernutrition far beyond its impact on hold had before a member or members became infected health services. HIV infection reduces household food with HIV, and lowers the quality of nutritional care received Trends and Issues in Child Undernutrition | 99 Figure 8.12 Female Adult Literacy Rates and Girls' Net Primary Enrollment Rates, by Country a. Adult female literacy rate, 2001 (percent) b. Girls' net primary school enrollment rate (percent) 30° 20° 10° 0° 10° 20° 30° 40° 50° 30° 20° 10° 0° 10° 20° 30° 40° IBRD 34525 MARCH 2006 Mediterranean Mediterranean Sea Sea 30° 30° 30° 30° Red Red Sea Sea 20° 20° 20° 20° 10° 10° 10° 10° 0° ATLANTIC 0° 0° ATLANTIC 0° OCEAN INDIAN OCEAN INDIAN OCEAN OCEAN 10° 10° 10° 10° more than 70 20° more than 60 20° 20° 20° 40 ­ 60 45 ­ 70 less than 40 less than 45 30° no data 30° 30° no data 30° international boundaries international boundaries This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° Source: Data from UNDP 2003. Note: a. Adult female 15 years and older. b. Enrollments are of girls of primary school age attending primary school. by children and others dependent on individuals infected however, cannot be expected to enjoy nutrition security, with the virus. in part because of the lower-quality care they receive from The final underlying determinant of child nutritional their relatively less-well-educated mothers. In spite of the status to consider is that of the quality of care children increases in food production shown for the countries of receive. It was highlighted earlier that an important proxy the Sahel in figure 8.9, without knowledgeable care, sus- for this determinant is the educational attainment and liter- tained reduction in child undernutrition there cannot acy levels of women, as mothers with more education are be ensured. more knowledgeable about the care their children need. Two Similarly, we can also examine actual child care practices, measures of this are provided in figure 8.12. The first, female in particular whether mothers practice exclusive breastfeed- adult literacy, reflects past broad knowledge building. The ing in the first four to six months of life of their children. second, the net enrollment rate for girls in primary school, Exclusive breastfeeding limits the exposure of children to reflects the degree to which knowledge, both general and disease while providing all the nutrients an infant requires more nutrition-focused knowledge, is being gained to for normal growth. As presented in figure 8.13, a mixed pic- enhance the nutritional status of future generations. Where ture is seen on the degree to which exclusive breastfeeding one finds low literacy and low enrollment of girls in school, is practiced across the countries in Sub-Saharan Africa. one should also expect relatively higher levels of undernu- Certainly the fact that so many infants in Sub-Saharan trition now and in the future, all things being equal. Africa are not exclusively breastfed in their early infancy The patterns seen across Sub-Saharan Africa are mixed. accounts for a significant portion of the high levels of Encouraging levels for both measures are seen in many undernutrition and under-five mortality observed in the countries of eastern and central Africa. If the educational region, as well as for the speed with which infants become curriculum in these countries provides basic nutritional undernourished after birth. and child care principles, we should expect that poor-quality Finally, economic growth, income, and poverty levels care will not be a critical factor contributing to undernutri- provide important insights into understanding the levels tion. Many children in the Sahelian countries in particular, of undernutrition seen in Sub-Saharan Africa. With little 100 | Todd Benson and Meera Shekar Figure 8.13 Proportion of Infants Receiving Only Breast Milk tently ranked among the worst on the continent. Conflict in the First Year of Life, Selected Countries exacerbates poverty and poor governance. These govern- 100 ments have been unable to provide basic public goods, which results in a consequent lack of access to food, care, 90 health services, and a healthy environment for their citizens. 80 Although in several of these countries, encouraging signs 70 have been seen since the early years of the twenty-first cen- 60 breastfed tury, it will take several years of such positive trends before 50 only children in these countries can be brought up in conditions 40 that will result in their enjoying a healthy and active life. infants % 30 It is more challenging to account for why most countries 20 in Africa are exhibiting only poor to fair progress in ensur- 10 ing the nutritional well-being of their children. The majority 0 of Africans live in these countries. Assessing their perform- a i ia Ethiopi ana ania Gh Mal anda abwe ance in light of the various factors that determine child Tanz Ug Zamb Zimb undernutrition in order to judge what changes they might country make can only be done case by case. Source: Data from ORC/Macro 2004, DHS survey. In those countries in which food availability is deficient, Note: Each bar represents a two-month interval during the first year of life. Missing bars indicate that no infants were being exclusively breadfed in those age ranges (e.g., 9 to 10 months, food production must be enhanced at the same time as trade 11 to 12 months). policies must be reexamined to allow a more reliable supply growth, the various African economies are generating little of food from the global market. Other countries may exhibit additional wealth. quite high levels of food available at an aggregate level but Consequently, income levels remain stagnant and pur- still have crippling rates of undernutrition. In these coun- chasing power will likely decline. The low GDP per capita tries, attention should be directed to issues of access to food and the meager existence experienced by significant seg- (the distribution of resources and consumption levels across ments of the population in Sub-Saharan Africa, are unlikely society and within households) and to the context within to change for some time to come. The impact of this on which the food is used (sanitation, health services, level of levels of undernutrition is obvious. Stagnant incomes imply knowledgeable care, and the broad range of related issues no increased access to the food required for a healthy diet. outlined earlier). Where HIV rates are high, the links Necessary public services for improved child nutrition-- between HIV and nutrition must be addressed. Moreover, public and curative health services, educational systems, the quality of policy making and the effective and responsi- support to agriculture, and infrastructure for marketing and ble implementation of those policies are important basic trade--will receive insufficient resources. The prospects are determinants of the degree to which undernutrition is effec- gloomy for sharp reductions in child undernutrition within tively addressed in any country in Africa, as elsewhere. the context of stagnant economies and pervasive poverty. Actions Needed to Eliminate Undernutrition DISCUSSION The challenge of child undernutrition in Africa is immense. Of all the developing regions of the world, it is only in Sub- The conceptual framework of child undernutrition pre- Saharan Africa that large increases in the number of under- sented in figure 8.1 highlights the multisectoral nature of nourished are anticipated--worse, in some regions of the the nutrition problem. It is usually not possible to point to a continent increases in the prevalence of undernourished single factor to account for high levels of undernutrition. children are also expected. These patterns are further exac- However, the high levels of undernutrition seen in those erbated by the continuing spread of HIV infection. Recent nations in Africa that have experienced conflict and the patterns of action by governments in Africa and their devel- absence of an effective central government in recent years is opment partners in addressing undernutrition must be not surprising. Burundi, the Democratic Republic of Congo, changed if the realization of these grim projections is to be Liberia, Somalia, and Sierra Leone, when assessed based on avoided. It is quite evident that efforts to date are not suffi- those nutritional measures for which data exist, are consis- cient to meet the challenge. Trends and Issues in Child Undernutrition | 101 National governments must lead several aspects of the · Significant increases in the resources allocated to preven- effort to improve child nutrition. They have a responsibility tive public health interventions directed to young chil- to establish the conditions and institutions necessary to dren and their mothers are necessary. Since the window enable children and their caregivers to access the basic of opportunity for nutrition interventions is small, these requirements for improved nutrition outcomes. efforts must be directed during this brief time--from the mother's pregnancy through the child's first two years · Sustained and broadly based economic growth is needed. of life. In addition, basic preventive and curative health To end hunger in Sub-Saharan Africa by 2050, it is esti- services need to be made more available to enable young mated that the region must attain an annual 3.5 percent children in Africa to cope better with the heavy disease average growth rate in per capita GDP (Runge et al. burdens they face--a disease burden that can be seen as 2003). In the past decade, however, only half a dozen both an outcome and a cause of their poor nutritional countries had growth rates above 2.5 percent. status. Although it has been shown that income growth The health sector is usually responsible for coordinat- will improve nutrition, the process is relatively slow. ing and leading direct nutrition intervention programs. Consequently, we cannot expect that income growth Such interventions, particularly those that address alone will be sufficient to improve nutrition outcomes micronutrient undernutrition and broad child survival at the rates that are required to achieve one of the targets concerns, have proved effective and efficient in improv- of the first Millennium Development Goal of cutting the ing the nutritional status of, in particular, preschool chil- prevalence of underweight children in half by 2015. dren and women of childbearing age. In assessing how Studies for Tanzania show that even under an optimistic best to address undernutrition, it is crucial not to under- scenario of 5.0 percent per capita GDP growth and using estimate the impact of such direct nutritional interven- a 0.51 elasticity between income and nutrition outcomes, tions. These programs are important investments for income growth alone will take up to 2026 to achieve the improved human welfare and economic productivity. MDG. For this MDG target to be achieved in Tanzania Their benefits can be seen most dramatically in efforts through income growth alone, a per capita growth of to reduce the prevalence of low birthweight babies. 5.0 percent and an elasticity of 0.85 would be needed, Behrman, Alderman, and Hoddinott (2004) present the a scenario that is totally unrealistic for any country in results of a cost-benefit analysis of programs of various Africa. With more realistic income growth estimates, the sorts that seek to improve fetal nutrition and increase potential for achieving the MDG are even further in the infant birthweight--prenatal care, in the broadest sense. time horizon. However, the results of simulations under- The direct costs of the programs that they examined taken for Kagera region in Tanzania that are presented in ranged from US$14 to US$100 per low birthweight birth table 8.4 show that if income growth is combined with averted. They calculated the present discounted value of direct nutrition interventions, the chances of achieving the benefits that accrue from averting a low birthweight the MDG become real. birth to amount to more than US$550. The benefits from Table 8.4 Projected Impact of Economic Growth and Direct Large-Scale Nutrition Intervention Programs on Stunting and Poverty, Kagera Region, Tanzania, 2015 (percent) Reduction in prevalence of stunted children Projected per capita income growth Reduction in No large-scale Intervention covering Intervention covering Intervention covering (1993­2015) income poverty nutrition intervention 10% of population 50% of population all of population 0.0 0 0 2 22 49 1.0 44 5 8 28 54 2.0 67 11 14 32 58 3.0 84 17 17 36 61 Source: Alderman et al. 2005. Note: Italicized values meet the MDG targets. Other factors contributing to reduced malnutrition include non-farm income greater than 25 percent of total income: 2 percent; passable roads throughout the year in every village: 5 percent; additional year of education to every father: 8 percent. 102 | Todd Benson and Meera Shekar improved fetal nutrition accrue from a range of sources, for seeing that attention is paid, human resources are including reduced costs of health care and productivity made available, or adequate funds are allocated to nutri- gains from the increased physical and mental abilities tion, it can easily be ignored or addressed only in an of the child through his or her life. The most important uncoordinated, piecemeal fashion. However, the issue of of these benefits is simply from the increased economic improving child nutrition is not unique in this regard. productivity potential of the child. Through such Although the cross-sectoral barriers to effectively interventions, the health sector has played an important addressing undernutrition may be particularly salient, role in the gains that have been achieved in addressing most development issues ultimately require coordinated undernutrition for human development in Sub-Saharan cross-sectoral action. Such coordination is fostered by Africa. However, these efforts must be duplicated many political leaders exercising political will. Child undernu- times over to successfully fight child undernutrition. trition is a high-priority development challenge in most · In addressing child undernutrition in Sub-Saharan African countries. Individual sectors can make important Africa, agriculture cannot be ignored. Growth in food contributions to facilitating the access of households to supplies has the dual effect of increasing the income of what they require to ensure the nutrition of their chil- the farming household and reducing the prices house- dren. Ultimately, however, the responsibility for imple- holds must pay to acquire food in the marketplace, both menting activities across sectors lies with the political of which contribute to better nutrition. leaders. These leaders must be champions for improving · Education is a critical input to good nutritional status, nutrition within their countries, and this commitment particularly for girls. The nutrition education messages must be sustained over time. that need to be learned are relatively simple--the com- ponents of a balanced diet and information on how Scale and Commitment locally available foods can be used to build balanced diets, the value of exclusive breastfeeding and appropriate Ultimately, the outcome of all efforts to improve child nutri- and timely complementary feeding for infants, the tion must be realized at the level of the individual child. As importance of prenatal care and regular monitoring of is clear from the UNICEF framework on the determinants child growth, sanitation and maintaining a healthy envi- of nutritional status, however, many, if not most, of the ronment, and the control of infant and childhood ill- determinants of nutritional status operate on much broader nesses, in particular. geographic scales than that of the individual. What does this Although it is not an obvious element of nutrition imply for policy making and planning, resource allocation, strategies, ensuring that girls are able to go to school and and program implementation to reduce food and nutrition attain their full educational potential is a critical compo- insecurity? nent of any longer-term effort to enhance nutrition in Several issues must be kept in mind. First, at the broader Africa. scales, heterogeneity exists in the key constraints to · A close link exists between successful improvements in improved child nutrition. Consequently, although broad child nutrition and increasing women's social access to strategies to enhance child nutrition can be developed, resources they can use to improve care and increase the these strategies must be context specific in their diversity and quantity of food provided the children under implementation. their care. Consequently, improving the level of equity Second, and emerging in part from the first point, efforts between men and women is good for child nutrition. aimed at sustainably reducing child undernutrition that · Regarding institutional and policy issues, it is important involve strong central government planning and control are to consider the particular characteristics of nutrition as a unlikely to succeed, especially in Sub-Saharan Africa, where subject for policy making and program implementation decentralization is taking root. Locally conceived and imple- by national government institutions. Nutrition is usually mented action is the primary manner in which the barriers neglected in the formulation of government and sectoral and constraints to such security can be removed. The role policies and strategies and in the allocation of govern- of the central government should be much broader and ment resources. As already emphasized, addressing child looser, consisting of giving broad general direction to local undernutrition requires action across sectors. With efforts and facilitating those efforts by allocating resources, no one sectoral ministry responsible for nutrition and, providing needed expertise, offering institutional support, consequently, no strong sectoral advocates responsible and the like. Although many African countries face severe Trends and Issues in Child Undernutrition | 103 capacity constraints at local levels, such a locally derived that many African nations have developed in the past five approach to address undernutrition should serve as the years and that serve as master development plans for many model for such efforts. are explicit on the importance of investing in nutrition to Finally, policy making and program planning have always reduce poverty and generate sustained economic growth. been guided by those whose voices are heard within the Similarly, for governments in Africa to address problems policy-making arena. Over the past 10 years, democratic of undernutrition effectively, cross-sectoral action is political frameworks have increased in number across Sub- required. It is unlikely that sustainable reductions in the Saharan Africa. The possible increase in attention paid to number of African children who are undernourished will the "will of the people" and the new political calculations be achieved if undernutrition is seen as a problem for the required of politicians within these new competitive systems health sector or the agricultural sector alone to address. may either improve the attention given to or draw attention Sectoral plans and strategies should be oriented toward away from nutritional considerations in policy making. objectives that result in improved nutrition, along with their Consequently, community-level political leaders and others other long-standing objectives. Advocates for improved who seek government support for nutrition-related inter- nutrition must engage in the higher-level policy processes ventions are unlikely to succeed without strategically engag- guiding the revisions of the poverty reduction strategy ing in the political processes of which they are a part. papers and sectoral strategy documents. The key message Increasing attention to significantly reduce child under- should be that the arrow of cause and effect between nutrition requires dedicated efforts to gain more commit- improved child nutrition and income and broader and sus- ment at all policy levels--decentralized, national, and tained economic growth runs both ways. Just as income global--and as an issue of broad public concern. To com- growth enhances nutrition, healthy, active, well-nourished pete successfully within a democratic political arena, the children are an important precondition for sustained future issue must be communicated effectively and understood growth in income. Nutrition concerns must be among the widely, its significance for the welfare of all members of primary components of such strategies. society recognized, and action catalyzed around proposed Moreover, regional and global policy initiatives can be solutions. Dedicated and sustained commitment is required. drawn upon in support of advocacy for improved nutrition. Ultimately, the effort should be to seek to establish the polit- Increasing food supplies and reducing hunger are central ical will and commitment to devote the necessary resources objectives of the Comprehensive Africa Agriculture to reduce child undernutrition. Although these children Development Programme of the New Partnership for have a basic human right to be properly nourished, efforts Africa's Development (NEPAD 2002). Although nutrition to make sure that they are must still be evaluated and find security requires more than increased food supplies, this support within the arena of politics. NEPAD initiative does commit African governments to addressing undernutrition. The Millennium Development Goals also imply the need Policies and Resources Needed to Support These Actions for strong resource allocations to reducing child undernu- If we agree that policy, when developed in a democratic and trition. The prevalence of underweight children is one of the transparent manner, defines the common good and serves indicators being used to track progress toward attaining the as a statement of how government will prioritize its actions MDG target of halving between 1990 and 2015 the propor- and its expenditures, then eliminating child undernutrition tion of people who suffer from hunger. Recent assessments must be central to government policies across Africa. show that only 6 of the almost 50 countries in Sub-Saharan Although the broad trends in the prevalence of stunting are Africa will meet this target. In 13 countries, prevalence rates slightly downward, the aggregate continental trend is not in 2015 relative to 1990 are expected to have increased reflected consistently at the individual country level as eco- (Chhabra and Rokx 2004). The trends in the prevalence of nomic, food, and social sector policies change or are not underweight children, shown in figure 8.14, points to just given consistent support, conflict shatters any progress that how large the gap from the goal will be if current patterns has been made, or emergency food crises are not effectively are maintained in all regions of Sub-Saharan Africa. Clearly, managed. Clarity in purpose and proper allocation of current levels of attention to undernutrition among chil- resources in line with appropriate policies is a necessary dren are grossly inadequate. The commitment to allocate element in any efforts to sharply reduce undernutrition. resources to attain the MDGs was made by African govern- Thus, it is critical that the poverty reduction strategy papers ments in the support they gave the United Nations 104 | Todd Benson and Meera Shekar Figure 8.14 Effect of Economic Growth on Attaining the underweight children is one of the indicators being used to track progress Millennium Development Goal of Reducing the Prevalence of toward attaining the target of halving between 1990 and 2015 the propor- tion of people who suffer from hunger under the first MDG of eradicating Underweight Preschool Children, Tanzania extreme poverty and hunger. 30 28 children 26 24 REFERENCES 22 Alderman, H., J. Hoddinott, and B. Kinsey. 2003. "Long-Term 20 underweight Consequences of Early Childhood Malnutrition." Food Consumption of 18 and Nutrition Division Discussion Paper 168, International Food 16 Policy Research Institute, Washington, DC. 14 Alderman, H., H. Hoogeveen, and C. Rossi. 2005. "Reducing Child prevalence 12 % Malnutrition in Tanzania--Combined Effects of Income Growth and 10 Program Interventions." Policy Research Working Paper 3567, World 1998 2001 2004 2007 2010 2013 2016 Bank, Washington, DC. year Behrman, J. R., H. Alderman, and J. Hoddinott. 2004. "Hunger and Trend needed to attain MDG Malnutrition." Paper for the Copenhagen Consensus: Challenges and 2.1% income growth and 0.30 elasticity; MDG in 2109 Opportunities series. Photocopy of January draft, University of 5.0% income growth and 0.30 elasticity; MDG in 2065 Pennsylvania, Philadelphia. 5.0% income growth and 0.51 elasticity; MDG in 2026 Black, R. E., S. S. Morris, and J. Bryce. 2003. "Where and Why Are 10 Million Children Dying Every Year?" Lancet 361: 2226­34. Source: Data from World Bank 2006. Chhabra, R., and C. Rokx. 2004. "The Nutrition MDG Indicator-- Note: Income growth is per capita. The elasticity indicates the magnitude of the effect of income growth on the reduction in underweight prevalence. Interpreting Progress." Health, Nutrition, and Population Discussion Paper, World Bank, Washington, DC. CIESIN (Center for International Earth Science Information Network). Millennium Declaration of 2000. Promoters of improved 2004. "Sub-National Estimates of Underweight Prevalence in Africa." Maps. Earth Institute of Columbia University, New York. nutrition in Africa must now hold the governments de Onis, M., M. Blössner, E. Borghi, R. Morris, and E. A. Frongillo. 2004. accountable to their pledge. "Methodology for Estimating Regional and Global Trends of Child Similarly, budgetary allocations by central governments Malnutrition." International Journal of Epidemiology 33 (6): 1260­70. should reflect the importance that improved nutrition has Ezzati, M., A. D. Lopez, A. Rodgers, S. Vander Hoorn, C. J. L. Murray, and the Comparative Risk Assessment Collaborating Group. 2003. for the welfare of all people and the immense economic "Selected Major Risk Factors and Global and Regional Burden of benefits it provides for relatively little cost. In this regard, Disease." Lancet 360: 1347­60. the donor funding that supports the vast majority of direct FAO (Food and Agriculture Organization [of the United Nations]). 2003. nutrition programs in Sub-Saharan Africa should be viewed The State of Food Insecurity in the World, 2003. Rome: FAO. as a secondary resource to complement government's own. Gwatkin, D. R., S. Rutstein, K. Johnson, R. B. Pande, and A. Wagstaff. 2003. Initial Country-Level Information about Socio-Economic Differences in African governments should make expenditure on nutrition Health, Nutrition, and Population. 2 vols. Washington, DC: World Bank. programs a core element of their annual budgets. Online analytical results: http://devdata.worldbank.org/hnpstats/. All African countries can eliminate child undernutrition. Horton, S. 1999. "Opportunities for Investments in Nutrition in Low- Income Asia." In Investing in Child Nutrition in Asia, ed. J. Hunt and What is needed is commitment to this goal, identification of M. G. Quibria, 246­73. Asian Development Nutrition and Develop- the right mix of strategies, dedicated efforts to marshal the ment Series 1. Manila: Asian Development Bank. human, institutional, and material resources necessary for Horton, S., and J. Ross. 2003. "The Economics of Iron Deficiency." Food the task, followed by the application of the political will to Policy 28: 51­75. undertake the actions at a scale necessary to achieve it. Jonsson, U. 1993. "Integrating Political and Economic Factors within Nutrition-Related Policy Research: An Economic Perspective." In The Political Economy of Food and Nutrition Policies, ed. P. Pinstrup- Andersen, 193­205. Baltimore: Johns Hopkins University Press. NEPAD (New Partnership for Africa's Development). 2002. Comprehen- NOTE sive Africa Agriculture Development Programme. Rome: FAO. ORC/Macro. 2004. "Demographic and Health Surveys STATcompiler." 1. The three measures indicate different nutritional processes. Children DHS data extraction tool. http://www.measuredhs.com/start.cfm. who are wasted have suffered acute nutritional deprivation in the recent Pelletier, D. L., and E. A. Frongillo Jr. 2003. "Changes in Child Survival Are past, such as occurs during famine. In contrast, stunted children have Strongly Associated with Changes in Malnutrition in Developing experienced cumulative retardation in their physical growth due to any of Countries." Journal of Nutrition 133: 107­19. a range of factors, diet-related or otherwise. The nutritional processes experienced by underweight children are nonspecific, possibly resulting Pelletier, D. L., E. A. Frongillo Jr., D. G. Schroeder, and J.-P. Habicht. 1994. both from chronic and from acute undernutrition. The prevalence of "A Methodology for Estimating the Contribution of Malnutrition to Trends and Issues in Child Undernutrition | 105 Child Mortality in Developing Countries." Journal of Nutrition 124 UNICEF (United Nations Children's Fund). 1990. Strategy for Improved (Suppl. 10): 2106S­122S. Nutrition of Children and Women in Developing Countries. A UNICEF Runge, C. F., B. Senauer, P. G. Pardey, and M. W. Rosegrant. 2003. Ending policy review. New York: UNICEF. Hunger in Our Lifetime: Food Security and Globalization." Baltimore: ------. 2003. "The State of the World's Children 2003." Database. Johns Hopkins University Press. http://www.unicef.org/sowc03. SCN (United Nations Standing Committee on Nutrition). 2000. Fourth UNICEF and MI (Micronutrient Initiative). 2004. "Vitamin and Mineral Report on the World Nutrition Situation. Geneva: SCN in collaboration Deficiency: A Global Damage Assessment Report." http://www. with International Food Policy Research Institute. micronutrient.org/reports/default.asp. ------. 2004. Fifth Report on the World Nutrition Situation: Nutrition for UNDP (United Nations Development Programme). 2003. "Human Improved Development Outcomes. Geneva: SCN. Development Indicators 2003." Database. New York: UNDP. http:// Smith, L. C., and L. Haddad. 2000. Explaining Child Malnutrition in hdr.undp.org/ Developing Countries: A Cross-Country Analysis. Research Report 111. Washington, DC: International Food Policy Research Institute. WHO (World Health Organization). 2003. "Global Database on Child Growth and Malnutrition." Online database. Geneva. http://www.who. Smith, L. C., U. Ramakrishnan, A. Ndiaye, L. Haddad, and R. Martorell. int/nutgrowthdb. 2003. The Importance of Women's Status for Child Nutrition in Developing Countries. Research Report 131. Washington, DC: World Bank. 2003. "The 2003 World Development Indicators CD-ROM." International Food Policy Research Institute. Washington DC: Development Data Group, World Bank. Svedberg, P. 1990. "Undernutrition in Sub-Saharan Africa: Is There a ------. 2006. Repositioning Nutrition as Central to Development: A Gender Bias?" Journal of Development Studies 26 (3): 469­86. Strategy for Large-Scale Action. Washington, DC: World Bank. 106 | Todd Benson and Meera Shekar Chapter 9 Diarrheal Diseases Cynthia Boschi-Pinto, Claudio F. Lanata, Walter Mendoza, and Demissie Habte Of the estimated total 10.6 million deaths among children Similarly to all-cause mortality, global estimates of the younger than five years of age worldwide, 42 percent occur number of deaths due to diarrhea have shown a steady in the World Health Organization (WHO) African region decline, from 4.6 million in the 1980s (Snyder and Merson (Bryce et al. 2005). Although mortality rates among these 1982) to 3.3 million in the 1990s (Bern et al. 1992) to children have declined globally from 146 per 1,000 in 1970 to 2.5 million in the year 2000 (Kosek, Bern, and Guerrant 79 per 1,000 in 2003 (WHO 2005), the situation in Africa is 2003). However, diarrheal diseases continue to be an impor- strikingly different. As compared with other regions of the tant cause of morbidity and mortality worldwide, and world, the African region shows the smallest reductions in despite all advances in health technology, improved man- mortality rates and the most marked slowing down trend agement, and increased use of oral rehydration therapy (figure 9.1). The under-five mortality rate in the African (ORT) in the past decades, they remain among the five region is seven times higher than that in the European region. major killers of children under five years of age. In 1980 this difference was equal to 4.3 times (WHO 2005). In contrast to mortality trends, morbidity due to diar- During the 1990s, the decline of under-five mortality rhea has not shown a parallel decline, and global estimates rates in 29 countries of the world stagnated, and in 14 coun- remain between two and three episodes of diarrhea per child tries rates went down but then increased again. Most of under five per year. Kosek, Bern, and Guerrant (2003) these countries are from the African region (WHO 2005). A estimated a global median incidence of diarrhea to be 3.2 factor that may contribute to this situation is the human episodes per child-year in the year 2000, similar to those immunodeficiency virus/acquired immune deficiency found in previous reviews by Snyder and Merson (1982) syndrome (HIV/AIDS) epidemic in the region, but an and by Bern and colleagues (1992) as well as to those underlying weakness of the implementation capacity of the reported in the first edition of Disease Control Priorities in health system is also likely to blame (Walker, Schwartländer, Developing Countries (Jamison et al. 1993). and Bryce 2002). 107 Figure 9.1 Slowing Progress in Child Mortality for each subregion whenever available data permitted. (per 1,000 births) Countries included in each subregion are listed in appendix 250 table 9A.1. 5 200 under DATA SOURCES AND LITERATURE REVIEW: 150 children MORBIDITY of 100 rate The usual sources of diarrhea morbidity data are either 50 national surveys, such as Demographic and Health Surveys mortality (DHSs) and the United Nations Children's Fund (UNICEF) 0 1970 1980 1990 2000 2003 Multiple Indicator Cluster Surveys (MICS, conducted from year 1996 to 2000; http://www.childinfo.org/MICS2/Gj99306k. Africa Western Pacific htm, accessed April 12, 2003), or the published literature. Eastern Mediterranean Western Pacific without China The main limitation of using currently available national World Americas Southeast Asia Europe survey data to estimate diarrhea morbidity is the cross- Southeast Asia without India sectional nature of data collection. The information obtained from these surveys is of diarrhea prevalence in the Source: Adapted from WHO 2005. two weeks previous to the survey, which does not account for seasonality. Therefore, data are not comparable either across This chapter reviews available information published sites or over time. Moreover, there is a potential for impor- since the 1980s on the morbidity and mortality burden of tant recall bias in such morbidity surveys (Boerma et al. diarrheal diseases in children under five years of age in the 1991; Snow et al. 1993). Some of the major limitations of WHO African region. longitudinal studies are lack of representativeness, possible site bias, low frequency of surveillance visits, and recall bias. Most reviews carried out so far (Bern et al. 1992; Kosek, METHODS Bern, and Guerrant 2003; Snyder and Merson 1982) have relied on published studies to estimate the incidence or Reliable and comparable estimates of morbidity and mor- prevalence of diarrheal disease. Some of the limitations of tality are difficult to obtain because of variations in method- this type of study are the small number of data points and ology, failure to standardize case definitions of diarrhea, the lack of representativeness, given the specific sites where and seasonal nature of the disease, among other factors. most studies are carried out. Nevertheless, such estimates, mainly based on the available The most recent morbidity review (Kosek, Bern, and studies in the literature, are provided here. Guerrant 2003) included five prospective studies from HIV infection has added considerably to the burden of African countries, carried out between 1987 and 1990: two diarrheal diseases among adults and children. This is of par- studies were from the AFR D subregion (Guinea-Bissau and ticular importance in African countries that show high HIV Nigeria) and three from the AFR E subregion (Democratic prevalence. However, the scarcity of data makes it difficult to Republic of Congo, Kenya, and Zimbabwe). These studies quantify comorbidity and its contribution to the mortality are listed in appendix table 9A.2. burden. Because of the probable influence of HIV/AIDS, especially in the mortality due to diarrheal disease, we have followed the WHO's division of the African region into DATA SOURCES AND LITERATURE REVIEW: two subregions, which takes into account mortality levels: MORTALITY the AFR D subregion (high child and high adult mortality) and the AFR E subregion (high child and very high adult Nationally representative surveys such as the DHS do not mortality). Stratum E includes the countries in Sub-Saharan usually report causes of death, but the number of diarrhea- Africa where HIV/AIDS has had a substantial impact associated deaths can be obtained from either vital statistics (Mathers et al. 2002). Specific estimates have been provided registration systems or from special study populations. 108 | Cynthia Boschi-Pinto, Claudio F. Lanata, Walter Mendoza, and Demissie Habte From each of these sources, the proportion of deaths attrib- Figure 9.2 Sites with Available Under-Five Diarrhea Mortality uted to diarrhea can be estimated as well as diarrhea Data mortality rates. However, the representativeness and accu- 30° 20° 10° 0° 10° 20° 30° 40° 50° IBRD 34526 MARCH 2006 racy of the data vary according to the type of source and var- Mediterranean Sea ious study design features. 30° 30° The main limitations of vital registration systems are underreporting of the number of deaths and miscoding of Red 20° Sea 20° the causes of death. Most of the limitations described for the use of longitudinal studies for estimating morbidity also 10° 10° apply to mortality estimation, such as lack of representative- ness, possible site bias, and misclassification of the causes of death. 0° ATLANTIC 0° OCEAN INDIAN The only countries in the African region for which there OCEAN is some reported vital registration (VR) coverage are 10° 10° Mauritius, South Africa, and Zimbabwe. The coverage for Mauritius was reported to be 100 percent in the year 2000. 20° 20° sites with under-five diarrhea Only 1.4 percent of all deaths among children under five mortality data international boundaries were due to diarrhea in this country. The latest informa- 30° 30° tion available for South Africa and Zimbabwe (1996 and 1990, respectively) reported estimated coverage rates of less This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. than 50 percent (http://www.who.int/whosis/mort/table4). 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° Therefore, in the African region, most information on Source: Authors. cause-specific mortality relies on special studies available in the literature. The studies included in our review were identified by illustrates the countries and the sites in each country for a systematic search of the scientific literature published which studies were available. since 1980, performed through the WHO's library on The 24 longitudinal studies identified in the current Medline/Pubmed using the following terms: Africa, mortal- literature represent an important increase in information ity, different spellings of "diarrhea," and all terms combined. when compared with the 7 available studies included in No restriction was placed on publication language. The the previous review (Kirkwood 1991). There has been more reference sections of these studies were then reviewed to than a threefold increase in the available literature of longitu- identify additional studies. dinal studies reporting diarrhea mortality in Sub-Saharan The review performed by Kirkwood (1991) for the previ- Africa. This increase in the number of publications from ous edition of Disease and Mortality in Sub-Saharan Africa Sub-Saharan Africa resulted in a doubled number of African included data from cross-sectional studies and from countries for which diarrhea mortality data were available, national diarrhea programs. In the current review we have thus adding to the precision of the current estimates. considered only longitudinal studies. Inclusion criteria were the following: studies carried out in countries from Sub- Saharan Africa, studies published from 1980 on, studies DATA SOURCE AND LITERATURE REVIEW: containing diarrhea-specific mortality data, studies contain- ETIOLOGY ing a minimum of 25 total deaths, community-based stud- ies with at least one year of follow-up, and a follow-up time Similarly to what was described for morbidity and mortality multiple of 12 months to minimize seasonal effects. estimates, the main data source to estimate diarrheal Twenty-four studies were identified that met the above etiology is the published literature. We conducted a search of criteria and were therefore included in this review (appendix papers published since 1990, using Medline/Pubmed; the table 9A.3). They were carried out in 15 (33 percent) of the terms used were as follows: Africa, Sub-Saharan Africa, 46 countries of the African region: seven were carried out diarrhea, etiology, epidemiology, and children. Different in countries from AFR D, and eight from AFR E. Figure 9.2 spellings as well as combinations of terms were also Diarrheal Diseases | 109 Table 9.1 Median Estimates of Episodes of Diarrhea per Child per Year in the African Region, by Age Group Age group Review 0­5 months 6­11 months 1 year 2 years 3 years 4 years Snyder and Merson 1982 2.6 4.3 2.3 1.0 0.7 0.4 Bern et al. 1992 2.7 4.5 2.4 2.7 1.9 1.7 Kosek, Bern, and Guerrant 2003 2.5 4.1 2.9a 2.2a -- -- Source: Authors. Note: -- not available. a. Only one observation available. considered. In addition, we considered keywords for specific Table 9.1 summarizes the results from these studies. Unlike etiologies known to cause diarrhea in children. Articles pub- mortality, estimates of morbidity due to diarrhea do not lished in languages other than English, Spanish, Portuguese, show a decline over time, according to the reviews carried Italian, and French that did not have an English abstract out. Estimates remain consistent for all age groups for which were not included. Further sources were identified from data are available. However, the number of observations, cross-references, consultation with experts in the field, and varying from three to five, is low for all three studies, and the use of the "related articles" link in PubMed. uncertainty that prevails because of this low number and the We used the following inclusion criteria: studies carried different sites where the studies were carried out should out at community level, among outpatient and inpatient be taken into account when interpreting these data. health services; studies that covered at least 12 months of Because some of the mortality studies were carried out in surveillance; and studies in which one or more causes of more than one country or more than one point in time, they diarrhea were identified through the use of standard labora- provided a total of 27 data points to be included in the tory procedures. Exclusion criteria were studies reporting analysis. More than 50 percent of these 27 data points diarrhea outbreaks, studies carried out among children with showed proportions of diarrhea mortality between 12 per- HIV/AIDS, studies reporting nosocomial infections, and cent and 17 percent; 19 out of the 27 data points (70 per- studies carried out in day-care centers. cent) had proportions between 12 percent and 19 percent; Thirty-four studies were identified that met the above cri- and only 8 data points provided proportions greater than teria, covering 12 African countries, 6 from each of the two 20 percent. Because of the skewness of the frequency distri- African subregions. These are listed in appendix table 9A.4. butions, we chose to use medians rather than means to calculate diarrhea proportional mortality for each African subregion. REVIEWS AND ESTIMATIONS We have used two different approaches to estimate the numbers and proportions of diarrhea deaths for Sub- In the 1980s Snyder and Merson (1982) reviewed 24 pub- Saharan Africa in the year 2000: calculation of simple lished studies in order to estimate morbidity and mortality medians for each African subregion and extrapolation from from diarrheal disease. Three of these studies, carried out in the regression of medians of diarrhea proportional mortality the African region (Ethiopia, Kenya, and Nigeria), reported against time. the annual number of episodes of diarrhea per child by age group. Simple Medians of Diarrhea Proportional Mortality In an attempt to update these estimates, Bern and colleagues (1992) reviewed articles published between 1980 As a first approach to estimating the number of deaths due and 1990. There were seven studies available for the African to diarrhea among children younger than five years of age in regions, covering three countries: The Gambia, Ghana, and the African region for the year 2000, we applied the median Nigeria. of the proportions of diarrhea deaths to the total number of Kosek, Bern, and Guerrant (2003) included four studies deaths among children under five in each of the two African from the African region in their review, which had been subregions. These medians were similar for the two subre- carried out in the Democratic Republic of Congo, Guinea- gions: 17.7 percent (interquartile intervals [IQI] 12.7­24.5 Bissau, Kenya, Nigeria, and Zimbabwe (appendix table 9A.2). percent) in AFR D and 17.6 percent (IQI 12.9­19.3 percent) 110 | Cynthia Boschi-Pinto, Claudio F. Lanata, Walter Mendoza, and Demissie Habte in AFR E. The WHO estimates that approximately 4.3 mil- Figure 9.4 Medians of Diarrheal Proportional Mortality lion children under five died in Africa in the year 2000: among Children under Five in the African Region and Other 2.0 million in AFR D and 2.3 million in AFR E (WHO, Developing Regions, 1980­95 unpublished data). Applying the proportion of diarrhea 30 deaths estimated to have occurred among children under five in Sub-Saharan Africa in the year 2000 to the total y 0.78x 1,571 number of deaths in these children in that same year yields 20 deaths an approximate total of 760,000 diarrheal deaths. diarrhea y 0.055x 126.1 10 % Proportions over Time It has been suggested that at least for some countries diar- 0 rhea mortality has been declining over the past years, mostly 1980 1985 1990 1995 year due to the spread of ORT use (Baltazar, Nadera, and Victora 2002; Miller and Hirschhorn 1995; Victora et al. 1996; African region other developing regions Victora et al. 2000). We thus examined the medians of the Source: Authors, from studies listed in appendix table 9A.3. proportions for studies carried out in Sub-Saharan Africa in different time periods. When comparing the median data points observed in the Because there were only two observations in the late African region with those from other developing regions of 1970s, we disregarded data from this time period and exam- the world (figure 9.4), we note that although there was a ined the following time periods: 1980­84, 1985­89, and steep decline in the proportions of diarrhea mortality in 1990­94. Figure 9.3 shows that there were virtually no the other regions ( 0.78), virtually no decline was changes in diarrhea proportional mortality ( 0.055) observed in Sub-Saharan Africa. These observations corre- in the African region over the years 1980­95. Indeed, the spond to an approximately 33 percent decline in the propor- medians of the proportion of diarrhea deaths were equal to tion of diarrheal deaths between 1980 and 1995 in the other 16.6 percent in the early 1980s, 17.7 percent in the late developing regions of the world and basically no changes in 1980s, and 16.1 percent in the early 1990s. If the situation the African region. remains the same and no major changes have occurred during more recent years, for which no data were available, Comparison of Estimates from Recent Reviews the proportion of diarrhea deaths in the year 2000 could be of the Burden of Diarrhea Mortality estimated to be equal to 16.2 percent. This corresponds to an estimated total of 700,000 deaths, similar to the estimate Kosek, Bern, and Guerrant (2003) recently reviewed 30 obtained with the simple median calculation. studies from all developing regions of the world that had available data on diarrhea proportional mortality for chil- dren under five years of age and that were published in the Figure 9.3 Medians of Diarrheal Proportional Mortality among Children under Five in the African Region, 1980­95 1990s. Ten of these studies were from the African region and covered seven different countries (five from AFR D and two 40 from AFR E). The medians of the diarrhea proportional y 0.055x 126.1 mortality were 23.2 percent in AFR D and 14.2 percent in 30 AFR E. By applying these medians to the total number of deaths deaths among children under five estimated by the WHO in 20 each of the two subregions in the year 2000, estimates of diarrhea 454,000 deaths for AFR D and 329,000 deaths for AFR E % 10 were obtained. These correspond to a total 783,000 deaths for the whole region. 0 Morris, Black, and Tomaskovic (2003) have also recently 1980 1985 1990 1995 2000 year reviewed the literature on causes of death among children under five and used a metaregression model with some Source: Authors, from studies listed in appendix table 9A.3. Note: extrapolation. selected covariates to estimate the proportional distribution Diarrheal Diseases | 111 Table 9.2 Estimated Number and Proportion of Deaths Due to Diarrhea among Children under Five (thousands) AFR D AFR E African region Approaches (no.) (%) (no.) (%) (no.) (%) Median of proportions 354 17.7 405 17.6 759 -- Regression on time -- -- -- -- 700 16.2 Kosek, Bern, and Guerrant 2003 454 23.2 329 14.2 783 -- Morris, Black, and Tomaskovic 2003 -- -- -- -- 935 21.9 WHO 2005 -- -- -- -- 741 17.0 Source: Authors, from studies listed in chapter and in appendix table 9A.3. Note: -- not available. of under-five deaths by cause in Sub-Saharan Africa and The locations of these studies were rarely representative South Asia. The authors predicted that 21.9 percent (95 per- of the entire country population, as they were usually con- cent CI, 15.5­28.2 percent) of all deaths of children up to ducted in populations that are either easy to access or have four years of age in Sub-Saharan Africa in the year 2000 atypical mortality rates. Furthermore, using studies from a were due to diarrhea, corresponding to a total of 935,000 few countries to predict distributions for many countries or deaths attributable to diarrhea. a region would require empirical external validation, which Table 9.2 summarizes the estimates obtained from the we were not able to perform because of the unavailability of main approaches used in this review and from the recent other sources of data, such as vital registration data for Sub- work performed by Kosek, Bern, and Guerrant (2003) and Saharan Africa or nationally representative surveys that Morris, Black, and Tomaskovic (2003) as well as from the included causes of death. However, we have tried to stratify WHO's most recent mean estimates for the period 2000­03 countries according to their mortality patterns, especially in (WHO 2005). Given the uncertainty of the estimates due to what concerns mortality due to HIV/AIDS, to minimize dis- the scarcity and limited representativeness of data, a reason- crepancies between them. Although the observations over able and plausible range for the numbers and proportions of time should be interpreted with caution because of the dif- deaths due to diarrhea among children under five in Africa ferent sites where the studies were conducted, this problem can be provided by summarizing different approaches from was minimized by including a reasonable number of studies various methodologies and their respective results. The two from different sites for each of the three time periods under estimates provided for AFR D were 354,000 (18 percent) observation, which should provide an average of the distri- and 454,000 deaths (23 percent) and those provided for bution of mortality in these sites. Also, it should be kept in AFR E were 329,000 (14 percent) and 405,000 deaths mind that these estimates were summarized as one single (18 percent). By reviewing the five available estimates for the observation over time for all of Sub-Saharan Africa, based on total African region, we might conclude that approximately a few sites from some countries (figure 9.2). As countries 750,000 deaths that occurred among children under five in often vary widely in many important socioeconomic and the year 2000 were due to diarrhea, with a range of estimates health aspects, summarizing data across countries of a that varied between 700,000 (WHO 2005) and 935,000 region may obscure important differences. deaths (Morris, Black, and Tomaskovic 2003). Our estimates Some studies of childhood deaths in developing coun- are similar to the 741,000 deaths estimated by WHO for the tries have shown that causes of death established using African region, and recently published in TheWorld Health verbal autopsy methods are not always consistent with Report (WHO 2005). diagnoses based on more complete clinical data (Kalter, Our estimates of deaths from diarrhea among children in Gray, and Black 1990; Mobley et al. 1996; Snow et al. 1992). Sub-Saharan Africa relied on published epidemiological However, the estimates of the cause-specific mortality studies using mostly verbal autopsy methods and thus have fraction (proportions of deaths attributable to one cause) limitations inherent in the type of data used, such as lack of resulting from verbal autopsy studies may not necessarily be representativeness and site bias, observations over time, inaccurate, if misclassification is random. misclassification of the causes of death, and comparability Finally, the studies reviewed used different case defini- of data from different studies. tions of diarrhea as a cause of death and different methods 112 | Cynthia Boschi-Pinto, Claudio F. Lanata, Walter Mendoza, and Demissie Habte for assigning them, both of which limit their comparability. Table 9.3 Available Etiology Data, by Country and Study Setting Any review that attempts to compile and summarize data Community Inpatient Outpatient from the published literature, especially those data that use Countries studies studies studies Total both standard and nonstandard verbal autopsy as the means AFR D of ascertaining cause of death, faces these limitations. Cameroon -- -- 1 1 However, the thorough literature search and the restrictive Gambia, The 1 -- -- 1 inclusion criteria used in this review, such as population- Ghana 1 -- 1 2 based studies with follow-up time a multiple of 12 months Guinea-Bissau 4 -- -- 4 and studies with at least 25 total deaths, ensured that the Madagascar -- 1 -- 1 studies used for the current estimates consisted of the most Nigeria -- -- 8 8 valid information available. AFR E Enteropathogens, such as rotavirus, entero-adherent Central African Republic 3 -- -- 3 pathogenic Escherichia coli (EAEC), and enterotoxigenic Ethiopia -- -- 1 1 Escherichia coli (ETEC), have been identified as important Kenya 1 1 3 5 pathogens in diarrheal diseases, and rotavirus, which causes South Africa 1 3 1 5 severe complications of diarrhea, has been found to be the Zambia -- 1 -- 1 most prominent cause of death in the world (Bern and Glass Zimbabwe -- -- 1 1 1994; Bishop 1994; Haffejee 1995). The severity of the etiological agent can be assessed by Source: Authors. Note: -- not available. See appendix table 9A.4 for the studies included. the setting in which it was most frequently isolated (com- munity, inpatients, or outpatients). Less severe agents would be more frequently found in community settings, whereas However, because of the scarcity of data available for the more severe ones should be more common in either outpa- African region we did not pursue these calculations. tients or (mainly) inpatients. Median proportions of diar- The studies included in the estimates of etiology by setting rheal episodes attributable to each major cause of diarrhea are listed in more detail in appendix table 9A.4. Table 9.3 from community studies could therefore be applied to esti- shows the countries for which data on etiology were avail- mates of diarrhea morbidity to obtain episodes of diarrheal able by setting (community, inpatients, or outpatients). diseases by cause, and median proportions from inpatient Tables 9.4 and 9.5 present the estimated median of the studies could be applied to estimates of diarrhea mortality proportions of etiological agents identified and correspon- to calculate the number of diarrheal deaths by etiology. ding IQI by study site for subregions AFR D and AFR E, Table 9.4 Median of the Proportions of Etiological Agents among Children under Five in the AFR D Subregion, by Study Site Community Inpatient Outpatient Etiological agents Median (IQ range) Median (IQ range) Median (IQ range) Salmonella 2.3 (2.0­2.6) 1.3 (1­13.3) -- -- Shigella sp. 14.5 (1.5­27.5) 1.7 (1.1­8.8) -- -- Campylobacter 5.9 (3.1­8.7) 17.7 (7.7­30.5) -- -- V. cholerae 0.6 (0.6­0.6) 1.9 (1.9­1.9) -- -- ETEC 8.5 (4.6­12.4) 1.2 (0.5­1.8) -- -- EAEC 4.2 (3.8­4.6) 21.4 (1.2­24.0) -- -- Rotavirus 6.2 (3.5­24.9) 19.6 (16.3­33.0) 12.3 (12.3­12.3) Giardia 15.1 (14.7­18.2) -- -- -- -- Cryptosporidium 6.5 (5.7­7.7) -- -- -- -- E. hystolitica 4.1 (0.0­7.0) -- -- -- -- Coinfection 15.8 (4.8­34.6) -- -- -- -- Unknown 16.4 -- 32.8 -- 11.4 -- Source: Authors, from studies listed in appendix table 9A.4. Note: -- not available. Diarrheal Diseases | 113 Table 9.5 Median of the Proportions of Etiological Agents among Children under Five in the AFR E Subregion, by Study Site Community Inpatient Outpatient Etiological agents Median (IQ range) Median (IQ range) Median (IQ range) Salmonella 1.9 (0.6­1.9) 7.4 (7.4­7.4) 2.5 (1.4­3.6) Shigella sp. 2.9 (2.0­3.8) 6.5 (6.5­6.5) 4.8 (3.6­6.0) Campylobacter 11.3 (3.5­11.7) 10.1 (4.8­15.3) 3.8 (2.2­5.3) V. cholerae 0.3 (0.3­0.3) 0.5 (0.5­0.5) -- -- ETEC 1.9 (1.6­3.1) 4.9 (4.9­4.9) -- -- EAEC 8.8 (8.6­9.7) 8.8 (8.8­8.8) 21.6 (15.6­27.6) Rotavirus 9.8 (6.3­15.5) 23.5 (19.3­31.1) 24.3 (20.8­32.3) Giardia 6.4 (0.9­10.0) 4.9 (4.9­4.9) -- -- Cryptosporidium 2.5 (2.5­2.5) -- -- -- -- E. hystolitica 3.3 (0.3­6.3) 7.8 (7.8­7.8) -- -- Coinfection 7.0 (7.0­7.0) -- -- -- -- Unknown 43.0 -- 23.9 -- 27.6 -- Source: Authors, from studies listed in appendix table 9A.3. Note: -- not available. respectively. In our review, Giardia lamblia was more (Daniels et al. 1990; Haggerty et al. 1994; LaFond 1995; frequently isolated among children with diarrhea in the MacDougall and McGahey 2003). These and many other community studies, whereas EAEC was mainly seen among factors, such as poor housing and crowding, are intrinsical- outpatients. Rotavirus was the etiological agent most ly associated with poverty. Furthermore, poverty usually frequently isolated in both inpatient and outpatient health limits access to health care and restricts appropriate and bal- services. Therefore, it is likely that rotavirus is the leading anced diets. Inequities in exposure and resistance add up to cause of mortality due to diarrhea in Africa, as has been inequities in coverage of available preventive interventions, observed in other parts of the world. Coinfection of various access to an appropriate health provider, and care, making agents was reported in 16 percent and 7 percent of poor children more likely to become sick than the better-off community-level studies in AFR D and AFR E, respectively. children (Victora et al. 2003). Because our review did not include publications of diarrhea Some studies have identified a few family characteristics outbreaks, the magnitude of Vibrio cholerae, and Shigella as protective factors. These are monogamy of the father, dysenteriae type 1 is underrepresented in this review. defined residential area (Vaahtera et al. 2000), having a private Our estimates of the distribution of diarrhea etiology kitchen, and being cared for by the mother (Oni, Schumann, among children in Sub-Saharan Africa relied on published and Oke 1991). These factors are of special importance in epidemiological studies. The same limitations reported for Sub-Saharan Africa, where the AIDS epidemic has led to an the morbidity and mortality estimates therefore apply to unprecedented number of orphans (about 12 million by the these etiological estimates. Moreover, very few studies were end of 2001) that is likely to more than double during this identified through the literature search for the African decade (Dabis and Ekpini 2002). region, and of those, not all have attempted to identify the A WHO report on global water supply provides worri- whole set of etiological agents. some figures of current and future scenarios for Africa (WHO 2000). Of all the regions in the world, the African region was the only one showing a decline in the proportion of the population that had access to sanitation between 1990 THE ROLE OF RISK FACTORS FOR DIARRHEAL and the year 2000 (figure 9.5). DISEASE IN THE AFRICAN REGION Approximately 50 percent (300 million individuals) of the African population have no access to safe water, and Broadly recognized risk factors for diarrheal diseases 66 percent (400 million individuals) lack access to hygienic include little or no access to safe water and sanitation, as sanitation. It is expected that by the year 2020 these figures well as poor hygiene and feces disposal practices at home will rise to 400 million and 500 million, respectively. 114 | Cynthia Boschi-Pinto, Claudio F. Lanata, Walter Mendoza, and Demissie Habte Figure 9.5 Change in Sanitation Coverage by Region, and poorer nutritional status than boys (Helen Keller 1990­2000 International 1994; Sundary 1986). In Sub-Saharan Africa, few prospective studies lasting at least one year report sex 110 99 100 100 differences at the community level (Perch et al. 2001). Some 91 90 83 prospective studies conducted in health facilities on outpa- 80 77 72 72 tients (Gomwalk et al. 1993; Gomwalk, Gosham, and Umoh 70 1990) and inpatients (Mpabalwani et al. 1995; Steele et al. 60 54 53 48 1998) show that boys are more likely than girls to be taken coverage 50 % 40 34 to the facility because of diarrhea (boy-girl ratios are 2 to 1 30 54 54 26 22 and 4 to 1, respectively). However, nationally representative 20 studies, such as the DHSs, conducted from 1987 to 2001 in 10 several Sub-Saharan African countries show no significant 0 aran ica d erica aibbean nd sex differences for health care­seeking behavior and treat- b-Sah Afr East andrica uth Asia So East Asia and Pacific /CIS anStates ment received, whether it is given at home or at a health Su MiddleNorth Af Latin Amthe Car CEE ltic Ba indust. countries facility. Findings are summarized in table 9.8. region The African region has been a target of diarrheal 1990 2000 epidemic outbreaks for several decades. One of the most dramatic manifestations of these outbreaks occurred in July Source: WHO 2000. 1994 among Rwandan refugees in Goma, Democratic Note: CEE Central and Eastern Europe; CIS Commonwealth of Independent States; no data available for 1990. Republic of Congo (formerly Zaire), when almost 50,000 refugees died during a diarrhea epidemic. Table 9.6 gives data on feces disposal practices at home in The African region has replaced the Indian subcontinent urban and rural regions of four African countries. Such prac- as the new home of V. cholerae. The seventh cholera tices have provided the rationale for more preventive inter- pandemic that originated in Asia reached Africa in the early ventions and are likely to explain the higher prevalence of 1970s. In 2001 there were more than 170,000 reported cases diarrheal disease in rural areas.As in most developing regions of cholera, which represented 94 percent of the globally of the world, African countries' poorest populations live in reported cases. From these, 2,590 people died. Nearly all rural areas. Table 9.7 presents the medians of the prevalence countries in Sub-Saharan Africa now regularly report cases of diarrheal disease according to urban and rural areas. of cholera. Table 9.9 shows the reported number of cases and deaths from cholera in the world and in the African SOME ADDITIONAL COMMENTS region between 1996 and 2001. These figures, however, need to be interpreted with caution, since countries that have There is an increasing concern that gender disparities might endemic cholera appear not to have notified the WHO of influence the distribution of ill-health and treatment, any cases of cholera. particularly in under-five girls. Some evidence suggests that Early reports of Vibrio parahaemolyticus in gastroenteri- girls in developing countries are prone to higher mortality tis cases in Africa (Utsalo, Eko, and Antia-Obong 1991) Table 9.6 Feces Disposal Practices at Home, by Urban and Rural Residences in Four African Countries (percent) Urban Rural Child Throw Bury Throw Throw Child Throw Bury Throw Throw Country (year of uses in toilet/ in the outside outside uses in toilet/ in the outside outside DHS) toilet latrine yard the dwelling the yard toilet latrine yard the dwelling the yard Benin (2002) 4.5 35.9 -- 22.6 6.8 1.2 6.2 -- 39.1 32.4 Malawi (2000) 8.2 80.3 0.5 0.5 9.6 7.5 69.8 3.2 2.7 7 Uganda (2000/2001) 13.5 75.7 1 3.9 -- 7.5 60.5 6 8.5 10.4 Zimbabwe (1999) 26.2 52.6 5.5 0.3 13.9 7.4 37.6 24.4 4.5 16.4 Source: Data from Demographic and Health Surveys 1987­2002. Note: -- not available. Diarrheal Diseases | 115 Table 9.7 Prevalence (Median) of Diarrheal Disease, the Two THE ROLE OF INTERVENTIONS TO CONTROL Weeks before Survey, by Urban and Rural Site of Residence DIARRHEAL DISEASES in AFR D and AFR E (percent) There is sufficient evidence that several interventions are African subregion Urban Rural effective in the prevention and treatment of diarrheal diseases (Jones et al. 2003). These interventions are exclu- AFR D 17.3 20.2 sive breastfeeding, complementary feeding, safe water, AFR E 14.9 19.1 good sanitation and hygiene, zinc and vitamin A African region 16.7 20.1 supplementation, ORT, and antibiotics for dysentery. It is Source: Data from Demographic and Health Surveys 1987­2002. estimated that these interventions could prevent 22 per- cent of deaths due to diarrhea (Jones et al. 2003). Most of have been confirmed to be associated with the O3:K6 these interventions are feasible for implementation in low- pandemic strain (Ansaruzzaman et al. 2004; Chowdhury income countries such as those in the African region; how- et al. 2000), documenting the extension of this pandemic ever, the capacity to deliver these important interventions into the region. S. dysenteriae serotype 1 has also been doc- effectively should be strengthened (Bryce et al. 2003). The umented to cause outbreaks of dysenteric diarrhea in sev- availability of safe and effective rotavirus vaccines (Ruiz- eral African countries (Birmingham et al. 1997; Guerin Palacios et al. 2006; Vesikari et al. 2006), introduced in et al. 2003; Malakooti et al. 1997), including in refugee several countries in Latin America in 2005 are likely to camps (Paquet et al. 1995). These organisms are important complement these interventions, if effectively delivered. causes of morbidity and mortality in diarrhea outbreaks However, the stability of diarrhea rates observed in all among stable communities and more so in displaced pop- reviews done since the 1980s shows that despite the reduc- ulations or those affected by catastrophic events. The wide- tion of diarrhea mortality, most likely through better case spread use of antibiotics across all regions has increased the management, very little has been done to prevent the prevalence of antibiotic resistance in most bacterial entero- transmission of diarrheal diseases. The progress toward pathogens, increasing the risk in the region of an outbreak better water and sanitation observed in other regions has of S. dysenteriae serotype 1 with multiresistant strains. not yielded a reduction of diarrhea morbidity, suggesting Table 9.8 Children under Five Taken to Health Facilities or Receiving Treatment for Diarrheal Disease in the Two Weeks before Survey, by Sex (percent) Taken to Using RHS at Using home Receiving Receiving Receiving no health facility antibiotics home remedy/other ORS packet increased fluids treatment Boys 28.8 14.1 13.7 39.9 25.6 36.0 17.4 Girls 29.2 11.8 12.1 38.2 25.2 37.6 17.3 Source: Data from Demographic and Health Surveys 1987­2001. Note: RHS recommended home solutions; ORS oral rehydration salt. Table 9.9 Cases and Deaths Due to V. cholerae Reported to WHO, 1996­2001 1996 1997 1998 1999 2000 2001 World Total no. of cases 143,349 147,425 293,121 254,310 137,071 184,311 Total no. of deaths 6,689 6,274 10,586 9,175 4,908 2,728 Case-fatality rate (%) 4.7 4.3 3.6 3.6 3.6 1.5 African region Total no. of cases 108,535 118,349 211,748 206,746 118,932 173,359 Total no. of deaths 6,216 5,853 9,856 8,728 4,610 2,590 Case-fatality rate (%) 5.7 4.9 4.1 4.2 3.9 1.5 Source: League of Nations 1996­2002. 116 | Cynthia Boschi-Pinto, Claudio F. Lanata, Walter Mendoza, and Demissie Habte that poor hygiene practices (Yeager et al. 1999) and the has been proved. Large reductions in child mortality could ingestion of contaminated food (Lanata 2003) may be the be achieved with their implementation. Therefore, careful most important factors and where preventive interven- planning and evaluation of interventions to control cases tions, like handwashing (Curtis and Cairncross 2003), and deaths due to diarrhea will be important if under-five should be promoted. mortality is to be reduced and goal four of the Millennium Development Goals--to reduce under-five mortality by two- thirds by 2015, from the base year 1990 (United Nations CONCLUSION 2000)--is to be achieved in the African region. Despite data limitations in estimating accurate numbers of diarrhea cases and deaths, it becomes clear from the results APPENDIX of this and other reviews that diarrheal disease remains an important cause of morbidity and mortality among Table 9A.1 Regional Reporting Categories for Global Burden children under five years of age in the African region. As of Disease 2000: WHO African Subregions opposed to the declining trends in the proportion of diar- WHO Mortality WHO rhea mortality in other developing regions of the world, vir- region stratum Member States tually no decline has been observed in the African region AFR D Algeria, Angola, Benin, Burkina Faso, since the early 1980s. Diarrheal diseases remain one of the Cameroon, Cape Verde, Chad, major killers of children under five, being responsible for Comoros, Equatorial Guinea, Gabon, The Gambia, Ghana, Guinea, about 750,000 of a total of 4.3 million deaths of African Guinea-Bissau, Liberia, Madagascar, children up to four years of age. Mali, Mauritania, Mauritius, Niger, More important than the precision in the numbers and Nigeria, São Tomé and Principe, in the exact contribution of each pathogen to diarrhea Senegal, Seychelles, Sierra Leone, Togo morbidity and mortality are the patterns and trends shown AFR E Botswana, Burundi, Central African in this review. The fact that almost 40 percent of all diarrhea Republic, Republic of Congo, Côte deaths in children under five worldwide occur in the African d'Ivoire, Democratic Republic of region is striking. The diarrhea mortality burden among Congo, Eritrea, Ethiopia, Kenya, children under five in Sub-Saharan Africa reveals the per- Lesotho, Malawi, Mozambique, Namibia, Rwanda, South Africa, sistent magnitude of this preventable and treatable disease Swaziland, Uganda, Tanzania, in the region. Zambia, Zimbabwe The efficacy of existing interventions to prevent or treat Source: WHO 2002. diarrheal diseases and to thereby reduce diarrhea mortality Table 9A.2 Main Characteristics of the Studies Included in the Morbidity Review Authors and year Region and No. of diarrhea episodes per of publication country Study period N child-year, by age group AFR D Oni, Schumann, and Oke 1991 Nigeria 1989­90 351 3.3 episodes (0­5 months) 4.1 episodes (6­11 months) 2.9 episodes (1 year) 2.2 episodes (2 years) Mølbak et al. 1997 Guinea-Bissau 1987­90 1,314 10.4 episodes (0­4 years) AFR E Mirza et al. 1997 Kenya 1989­90 920 3.5 episodes (0­3 years) Manun'ebo et al. 1994 Dem. Rep. of Congo 1987­88 1,914 6.3 episodes (0­4 years) Moy et al. 1991 Zimbabwe 1987­88 204 1.8 episodes (0­5 months) 4.8 episodes (6­11 months) Source: Adapted from Kosek, Bern, and Guerrant 2003. Diarrheal Diseases | 117 Table 9A.3 Main Characteristics of the Studies Included in the Mortality Review Mid-year No. of Under-five % diarrhea Authors and year of publication Region and country of studya deaths mortality rateb deaths AFR D Bendib, Dekkar, and Lamdjadani 1993 Algeria 1987 1,502 -- 8.1 De Francisco et al. 1993 Gambia, The 1989 856 34.5 18.0 Greenwood et al. 1987 Gambia, The 1982 184 72.1 13.6c Greenwood et al. 1990 Gambia, The 1983 76 70.2 15.8 Greenwood et al. 1990 Gambia, The 1985 187 58.6 26.7 Jaffar et al. 1997 Gambia, The 1991 3,776 98.9 8.4 Mølbak et al. 1992 Guinea-Bissau 1988 153 55.6 30.7 Becker, Diop, and Thornton 1993 Liberia 1984 -- 100.0 13.7c Becker, Diop, and Thornton 1993 Liberia 1987 -- 79.0 12.1c Bradley and Gilles 1984 Nigeria 1978 151 -- 24.5 Ekanem, Asindi, and Okoi 1994 Nigeria 1991 314 -- 12.1 Jinadu et al. 1991 Nigeria 1990 120 62.2 32.5 Fontaine et al. 1984 Senegal 1983 42 21.6 26.2 Pison et al. 1993 Senegal 1987 75 17.3 20.3 Victora et al. 1993 Senegal 1986 1,517 63.1 35.0 Amin 1996 Sierra Leone 1990 559 33.0 12.7 Hodges and Williams 1998 Sierra Leone 1989 4,264 -- 17.7 AFR E Delacollette and Barutwanayo 1993 Burundi 1990 160 41.9 19.4 Georges et al. 1987 Central African Republic 1983 188 28.5 19.1 Delacollette et al. 1989 Congo, Dem. Rep. of 1986 358 69.0 8.4 Shamebo et al. 1991 Ethiopia 1988 436 49.0 8.4 Omondi-Odhiambo, Ginneken, and Kenya 1977 338 -- 19.5 Voorhoeve 1990 Kahn et al. 1999 South Africa 1994 216 -- 20.4 Mtango and Neuvians 1986 Tanzania 1984 325 40.1 16.9 Mtango and Neuvians 1986 Tanzania 1985 347 35.0 12.7 Mtango et al. 1992 Tanzania 1987 610 30.8 13.3 Watts, Ng'andu, and Wray 1990 Zambia -- 26 -- 18.2c Source: Data from sources in table. Note: -- not available. a. Some studies did not report mid-year of study and those have been either informed by contacting authors or approximately estimated from other available information in the study. b. Under-five mortality rate is the number of total under-five deaths per 1,000 children per year (obtained either directly from the study or calculated from data available in the study). c. Proportions corrected or adjusted to age group 0­59 months. 118 | Cynthia Boschi-Pinto, Claudio F. Lanata, Walter Mendoza, and Demissie Habte Table 9A.4 Main Characteristics of the Studies Included in the Etiology Review Authors and year of publication Region and country Study setting AFR D Koulla-Shiro, Loe, and Ekoe 1995 Cameroon Outpatient Rowland et al. 1985 Gambia, The Community Armah et al. 1994 Ghana Outpatient Nakano et al. 1990 Ghana Community Fisher et al. 2000 Guinea-Bissau Community Mølbak 1990 Guinea-Bissau Community Mølbak et al. 1994 Guinea-Bissau Community Perch et al. 2001 Guinea-Bissau Community Cassel-Beraud, Michel, and Madagascar Inpatient Garbarg-Chenon 1993 Akinyemi et al. 1998 Nigeria Outpatient Audu et al. 2002 Nigeria Outpatient Gomwalk, Gosham, and Umoh 1990 Nigeria Outpatient Gomwalk et al. 1993 Nigeria Outpatient Obi et al. 1997 Nigeria Outpatient Okeke et al. 2000 Nigeria Outpatient Pennap et al. 2002 Nigeria Outpatient Pennap et al. 2000 Nigeria Outpatient AFR E Georges-Courbot et al. 1987 Central African Republic Community Georges-Courbot et al. 1990 Central African Republic Community Georges-Courbot et al. 1988 Central African Republic Community Gedlu and Aseffa 1996 Ethiopia Outpatient Chunge et al. 1989 Kenya Community Mutanda et al. 1986 Kenya Outpatient Mutanda et al. 1985 Kenya Inpatient Nakata et al. 1999 Kenya Outpatient Saidi et al. 1997 Kenya Outpatient Griffiths, Steele, and Alexander 1992 South Africa Inpatient Mnisi, Williams, and Steele 1992 South Africa Inpatient Sebata and Steele 2001 South Africa Outpatient Steele et al. 1998 South Africa Inpatient Steele et al. 1988 South Africa Inpatient Mpabalwani et al. 1995 Zambia Inpatient Tswana et al. 1990 Zimbabwe Outpatient Source: Information from sources in table. Diarrheal Diseases | 119 REFERENCES Chunge, R. N., I. A. Wamola, S. N. Kinoti, J. Muttunga, L. N. Mutanda, N. Nagelkerke, L. Muthami, E. Muniu, J. M. Simwa, P. N. Karumba, and Akinyemi, K. O., A. O. Oyefolu, B. Opere, V. A. Otunba-Payne, and A. O. P. Kabiru. 1989. "Mixed Infections in Childhood Diarrhoea: Results of Oworu. 1998. "Escherichia Coli in Patients with Acute Gastroenteritis a Community Study in Kiambu District, Kenya." East African Medical in Lagos, Nigeria." East African Medical Journal 75: 512­15. Journal 66: 715­23. Amin, R. 1996. "Immunization Coverage and Child Mortality in Two Curtis, V., and S. Cairncross. 2003. "Effect of Washing Hands with Soap Rural Districts of Sierra Leone." Social Science and Medicine 42: on Diarrhoea Risk in the Community: A Systematic Review." Lancet 1599­1604. Infectious Diseases 3: 275­81. Ansaruzzaman, M., N. A. Bhuiyan, G. B. Nair, D. A. Sack, M. Lucas, J. L. Dabis, F., and E. R. Ekpini. 2002. "HIV-1/AIDS and Maternal and Child Deen, J. Ampuero, and C. L. Chaignat. 2004. The Mozambique Cholera Health in Africa." Lancet 359: 2097­104. Vaccine Demonstration Project Coordination Group. "Cholera in Daniels, D. L., S. N. Cousens, L. N. Makoae, and R. G. Feachem. 1990. "A Mozambique, Variant of Vibrio cholerae." Emerging Infectious Diseases Case-Control Study of the Impact of Improved Sanitation on Diarrhea 10: 2057­59. Morbidity in Lesotho." Bulletin of the World Health Organization 68: Armah, G. E., J. A. A. Mingle, A. K. Dodoo, A. Anyanful, R. Antwi, 455­63. J. Commey, and F. K. Nkrumah. 1994. "Seasonality of Rotavirus De Francisco, A., A. J. Hall, J. R. Schellenberg, A. M. Greenwood, and Infections in Ghana." Annals of Tropical Paediatrics 14: 223­30. B. M. Greenwood. 1993. "The Pattern of Infant and Childhood Audu, R., S. A. Omilabu, M. de Beer, I. Peenze, and A. D. Steele. 2002. Mortality in Upper River Division, The Gambia." Annals of Tropical "Diversity of Human Rotavirus VP6, VP7, and VP4 in Lagos State, Paediatrics 13: 345­52. Nigeria." Journal of Health, Population, and Nutrition 20: 59­64. Delacollette, C., and M. Barutwanayo. 1993. "Mortalité et morbidité aux Baltazar, J. C., D. P. Nadera, and C. G. Victora. 2002. "Evaluation of the jeunes âges dans une région à paludisme hyperendémique stable, National Control of Diarrhoeal Disease Programme in the Philippines, commune de Nyanza-Lac, Imbo Sud, Burundi." Bulletin de la Societé de 1980­93." Bulletin of the World Health Organization 80: 637­43. Pathologie Exotique 86: 373­79. Becker, S. R., F. Diop, and J. N. Thornton. 1993. "Infant and Child Delacollette, C., P. Van der Stuyft, K. Molima, C. Delacollette-Lebrun, Mortality Estimates in Two Counties of Liberia; Results of a Survey in and M. Wery. 1989. "Étude de la mortalité globale et de la mortalité liée 1988 and Trends Since 1984." International Journal of Epidemiology 22 au paludisme dans le Kivu montagneux, Zaire." Revue d' Epidémiologie (Suppl. 1): S56­63. et de Santé Publique 37: 161­66. Bendib, A., N. Dekkar, and N. Lamdjadani. 1993. "Facteurs associés à la Ekanem, E. E., A. A. Asindi, and O. U. Okoi. 1994. "Community-Based mortalité juvénile, infantile et néonatale. Résultats d'une enquête Surveillance of Paediatric Deaths in Cross River State, Nigeria." nationale en Algérie." Archives Françaises de Pediatrie 50: 741­47. Tropical and Geographical Medicine 46: 305­8. Bern, C., and R. I. Glass. 1994. "Impact of Diarrheal Diseases Worldwide." Fisher, T. K., H. Steinsland, K. Mølbak, R. Ca, J. R. Gentsch, P. Valentiner- In Viral Infections of the Gastrointestinal Tract, 2nd ed., ed. A. Z. Branth, P. Aaby, and H. Sommerfelt. 2000. "Genotype Profiles of Kapikian. New York: Marcel Dekker. Rotavirus Strains from Children in a Suburban Community in Bern, C., J. Martines, I. de Zoysa, and R. I. Glass. 1992. "The Magnitude Guinea-Bissau, Western Africa." Journal of Clinical Microbiology 38: of the Global Problem of Diarrhoeal Disease: A Ten-Year Update." 264­67. Bulletin of the World Health Organization 70: 705­14. Fontaine, O., B. Diop, J. P. Beau, A. Briend, and M. Ndiaye. 1984. "La Birmingham, M. E., L. A. Lee, M. Ntakibirora, F. Bizimana, and M. S. diarrhée infantile au Senegal." Médecine Tropicale: Revue du Corps de Deming. 1997. "A Household Survey of Dysentery in Burundi: Santé Colonial 44: 27­31. Implications for the Current Pandemic in Sub-Saharan Africa." Gedlu, E., and A. Aseffa. 1996. "Campylobacter Enteritis among Children Bulletin of the World Health Organization 75: 45­53. in North-West Ethiopia: A 1-Year Prospective Study." Annals of Tropical Bishop, R. 1994. "Natural History of Human Rotavirus Infection." In Viral Paediatrics 16: 207­12. Infections of the Gastrointestinal Tract, 2nd ed., ed. A. Z. Kapikian. Georges, M. C., C. Roure, R. V. Tauxe, D. M. Y. Meunier, M. Merlin, J. Testa, New York: Marcel Dekker. C. Baya, J. Limbassa, and A. J. Georges. 1987. "Diarrheal Morbidity Boerma, J. T., R. E. Black, A. E. Sommerfelt, S. O. Rustein, and G. T. Bicego. and Mortality in Children in the Central African Republic." American 1991. "Accuracy and Completeness of Mothers' Recall of Diarrhoea Journal of Tropical Medicine and Hygiene 36: 598­602. Occurrence in Pre-School Children in Demographic and Health Georges-Courbot, M. C., A. M. Beraud-Cassel, I. Gouandjika, and A. J. Surveys." International Journal of Epidemiology 20: 1073­80. Georges. 1987. "Prospective Study of Enteric Campylobacter Infections Bradley, A. K., and H. M. Gilles. 1984. "Pointers to Causes of Death in the in Children from Birth to Six Months in the Central African Republic." Malumfashi Area, Northern Nigeria." Annals of Tropical Medicine and Journal of Clinical Microbiology 25: 836­39. Parasitology 78: 265­71. Georges-Coubot, M. C., A. M. Cassel-Beraud, I. Gouandjika, J. Monges, Bryce, J., C. Boschi-Pinto, K. Shibuya, R. E. Black, and the Child Health and A. J. Georges. 1990. "A Cohort Study of Enteric Campylobacter Epidemiology Reference Group. 2005. "WHO Estimates of the Causes Infection in Children from Birth to Two Years in Bangui (Central of Death in Children." Lancet 365: 1147­52. African Republic)." Transactions of the Royal Society of Tropical Medicine and Hygiene 84: 122­25. Bryce, J., S. el Arifeen, G. Pariyo, C. Lanata, D. Gwatkin, and J. P. Habicht. 2003. Multi-Country Evaluation of IMCI Study Group. "Reducing Georges-Courbot, M. C., J. Monges, A. M. Beraud-Cassel, I. Gouandjika, Child Mortality: Can Public Health Deliver?" Lancet 362: 159­64. and A. J. Georges. 1988. "Prospective Longitudinal Study of Rotavirus Infections in Children from Birth to Two Years of Age in Central Cassel-Beraud, A. M., P. Michel, and A. Garbarg-Chenon. 1993. Africa." Annales de l'Institut Pasteur. Virology 139: 421­28. "Epidemiological Study of Infantile Rotavirus Diarrhea in Tananarive (Madagascar)." Journal of Diarrhoeal Diseases Research 11: 82­87. Gomwalk, N. E., L. T. Gosham, and U. J. Umoh. 1990. "Rotavirus Gastroenteritis in Pediatric Diarrhea in Jos, Nigeria." Journal of Chowdhury, N. R., S. Chakraborty, T. Ramamurthy, M. Nishibuchi, Tropical Pediatrics 36: 52­55. S. Yamasaki, Y. Takeda, and G. B. Nair. 2000. "Molecular Evidence of Clonal Vibrio Parahaemolyticus Pandemic Strains." Emerging Infectious Gomwalk, N. E., U. J. Umoh, L. T. Gosham, and A. A. Ahmad. 1993. Diseases 6: 631­36. "Influence of Climatic Factors on Rotavirus Infection among Children 120 | Cynthia Boschi-Pinto, Claudio F. Lanata, Walter Mendoza, and Demissie Habte with Acute Gastroenteritis in Zaria, Northern Nigeria." Journal of LaFond, A. K. 1995. "A Review of Sanitation Program Evaluation in Tropical Pediatrics 39: 293­97. Developing Countries." Activity Report 5, Environmental Health Greenwood, B. M., A. K. Bradley, P. Byass, A. M. Greenwood, A. Menon, Project, Arlington, VA. R. W. Snow, R. J. Hayes, and A. B. Hatib-N'Jie. 1990. "Evaluation of Lanata, C. F. 2003. "Studies of Food Hygiene and Diarrhoeal Disease." a Primary Health Care Programme in The Gambia. II. Its Impact on International Journal of Environmental Health Research 13: S175­83. Mortality and Morbidity in Young Children." Journal of Tropical League of Nations, Health Section of the Secretariat. 1996­2002. Weekly Medicine and Hygiene 93: 87­97. Epidemiological Record, vols 71­77. http://www.who.int/wer/ Greenwood, B. M., A. M. Greenwood, S. Bradley, S. Tulloch, R. Hayes, archives/en/ and F. S. J. Oldfield. 1987. "Deaths in Infancy and Early Childhood in MacDougall, J., and C. McGahey. 2003. "Three Community-Based a Well-Vaccinated, Rural, West African Population." Annals of Tropical Environmental Sanitation and Hygiene Projects Conducted in the Paediatrics 7: 91­99. Democratic Republic of Congo." Activity Report 119, USAID, Griffiths, F. H., A. D. Steele, and J. J. Alexander. 1992. "The Molecular New York. Epidemiology of Rotavirus-Associated Gastro-Enteritis in the Malakooti, M. A., J. Alaii, G. D. Shanks, and P. A. Phillips-Howard. 1997. Transkei, Southern Africa." Annals of Tropical Paediatrics 12: 259­64. "Epidemic Dysentery in Western Kenya." Transactions of the Royal Guerin, P. J., C. Brasher, E. Baron, D. Mic, F. Grimont, M. Ryan, P. Society of Tropical Medicine and Hygiene 91: 541­43. Aavitsland, and D. Legros. 2003. "Shigella Dysenteriae Serotype 1 in Manun'ebo, M. N., P. A. Haggerty, M. Kalengaie, A. Ashworth, and B. R. West Africa: Intervention Strategy for an Outbreak in Sierra Leone." Kirkwood. 1994. "Influence of Demographic, Socioeconomic and Lancet 362: 705­6. Environmental Variables on Childhood Diarrhea in a Rural Area of Haffejee, I. E. 1995. "The Epidemiology of Rotavirus Infections: A Global Zaire." Journal of Tropical Medicine and Hygiene 97 (1): 31­38. Perspective." Journal of Pediatric Gastroenterology and Nutrition 20: Mathers, C. D., C. Stein, D. Ma Fat, C. Rao, M. Inoue, N. Tomijima, C. 275­86. Bernard, A. Lopez, and C. J. L. Murray. 2002."Global Burden of Disease Haggerty, P. A., K. Muladi, B. R. Kirkwood, A. Ashworth, and M. 2000: Version 2 Methods and Results." Global Programme on Evidence Manunebo. 1994. "Community-Based Hygiene Education to Reduce for Health Policy Discussion Paper 50, WHO, Geneva. Diarrhoeal Disease in Rural Zaire: Impact of the Intervention on Miller, P., and N. Hirschhorn. 1995. "The Effect of a National Control of Diarrhoeal Morbidity." International Journal of Epidemiolology 23: Diarrheal Diseases Program on Mortality: The Case of Egypt." Social 1050­59. Science and Medicine 40: S1­30. Helen Keller International. 1994. "Summary Report on the Nutritional Mirza, N. M., L. E. Caulfield, R. E. Black, and W. M. Macharia. 1997. "Risk Impact of Sex-biased Behavior." Nutritional Surveillance Project, Factors for Diarrheal Duration." American Journal of Epidemiology 146 Dhaka, Bangladesh. (9): 776­85. Hodges, M., and R. A. M. Williams. 1998. "Registered Infant and Under- Mnisi,Y. N., M. M.Williams, and A. D. Steele. 1992."Subgroup and Serotype Five Deaths in Freetown, Siera Leone from 1987­1991 and a Compari- Epidemiology of Human Rotaviruses Recovered at Ga-Rankuwa, son with 1969­1979." West African Journal of Medicine 17: 95­98. Southern Africa." Central African Journal of Medicine 38: 221­25. Jaffar, S., A. Leach, A. M. Greenwood, A. Jepson, O. Muller, M. O. C. Ota, Mobley, C. C., J. T. Boerma, S. Titus, B. Lohrke, K. Shangula, and R. E. K. Bojang, S. Obaro, and B. M. Greenwood. 1997. "Changes in the Black. 1996. "Validation Study of a Verbal Autopsy Method for Causes Pattern of Infant and Childhood Mortality in Upper River Division, of Childhood Mortality in Namibia." Journal of Tropical Pediatrics 42: The Gambia, from 1989 to 1993." Tropical Medicine & International 365­69. Health 2: 28­37. Mølbak, K., P. Aaby, L. Ingholt, N. Hojlyng, A. Gottschau, and H. Jamison, D. T., H. W. Mosley, A. R. Measham, and J. L. Bobadilla. 1993. Andersen. 1992. "Persistent and Acute Diarrhea as the Leading Causes Disease Control Priorities in Developing Countries. New York: Oxford of Child Mortality in Urban Guinea-Bissau." Transactions of the Royal University Press. Society of Tropical Medicine and Hygiene 86: 216­20. Jinadu, M. K., S. O. Olusi, J. I. Agun, and A. K. Fabiyi. 1991. "Childhood Mølbak, K., N. Hojlyng, L. Ingholt, A. P. Da Silva, S. Jepsen, and P. Aaby. Diarrhea in Rural Nigeria. I. Studies on Prevalence, Mortality and Socio- 1990. "An Epidemic Outbreak of Cryptosporidiosis: A Prospective Environmental Factors." Journal of Diarrhoeal Diseases Research 9: Community Study from Guinea Bissau." Pediatric Infectious Disease 323­27. Journal 9: 566­70. Jones, G., R. W. Steketee, R. E. Black, Z. A. Bhutta, S. S. Morris, and the Mølbak, K., H. Jensen, L. Ingholt, and P. Aaby. 1997. "Risk Factors for Bellagio Child Survival Study. 2003. "How Many Child Deaths Can Diarrheal Disease Incidence in Early Childhood: A Community We Prevent This Year?" Lancet 362: 65­71. Cohort Study from Guinea-Bissau." American Journal of Epidemiology 146: 273­82. Kahn, K., S. M. Tollman, M. Garenne, J. S. Gear. 1999. "Who Dies from What? Determining Cause of Death in South Africa's Rural North- Mølbak, K., N. Wested, N. Hojlyng, F. Scheutz, A. Gottschau, P. Aaby, East." Tropical Medicine & International Health 4: 433­41. and A. P. J. da Silva. 1994. "The Etiology of Early Childhood Diarrhea: A Community Study from Guinea-Bissau." Journal of Infectious Disease Kalter, H. D., R. H. Gray, and R. E. Black. 1990. "Validation of Postmortem 169: 581­87. Interviews to Ascertain Selected Causes of Death in Children." Morris, S. S., R. E. Black, and L. Tomaskovic. 2003. "Predicting the International Journal of Epidemiology 19: 380­86. Distribution of Under-Five Deaths by Cause in Countries without Kirkwood, B. 1991. "Diarrhea." In Disease and Mortality in Sub-Saharan Adequate Vital Registration Systems." International Journal of Africa, ed. R. G. Feachem and D. T. Jamison, 134­57. New York: Oxford Epidemiology 32: 1041­51. University Press. Moy R. J., I. W. Booth, R. G. Choto, and A. S. McNeish. 1991. "Recurrent Kosek, M., C. Bern, and R. Guerrant. 2003. "The Global Burden of and Persistent Diarrhea in a Rural Zimbabwean Community: A Diarrheal Disease, As Estimated from Studies Published Between 1992 Prospective Study." Journal of Tropical Pediatrics 37: 293­99. and 2000." Bulletin of the World Health Organization 81: 197­204. Mpabalwani, M., H. Oshitani, F. Kasolo, K. Mizuta, N. Luo, N. Koulla-Shiro, S., C. Loe, and T. Ekoe. 1995. "Prevalence of Campylobacter Matsubayashi, G. Bhat, H. Suzuki, and Y. Numazaki. 1995. "Rotavirus Enteritis in Children from Yaounde (Cameroon)." Central African Gastro-Enteritis in Hospitalized Children with Acute Diarrhea in Journal of Medicine 4: 91­94. Zambia." Annals of Tropical Paediatrics 15: 39­43. Diarrheal Diseases | 121 Mtango, F. D. E., and D. Neuvians. 1986. "Acute Respiratory Infections in Attenuated Vaccine against Severe Rotavirus Gastroenteritis." New Children under Five Years. Control Project in Bagamoyo District, England Journal of Medicine 354: 11­22. Tanzania." Transactions of the Royal Society of Tropical Medicine and Saidi, S. M., Y. Iijima, W. K. Sang, A. K. Mwangudza, J. O. Oundo, K. Taga, Hygiene 80: 851­58. M. Aihara, K. Nagayama, H. Yamamoto, P. G. Waiyaki, and T. Honda. Mtango, F. D. E., D. Neuvians, C. V. Broome, A. W. Hightower, and A. Pio. 1997. "Epidemiological Study on Infectious Diarrhoeal Diseases in 1992. "Risk Factors for Deaths in Children under 5 Years Old in Children in a Coastal Rural Area of Kenya." Microbiology and Bangamoyo District, Tanzania." Tropical Medicine and Parasitology 43: Immunology 41: 773­78. 229­33. Sebata, T., and D. Steele. 2001. "Atypical Rotavirus Identified from Young Mutanda, L. N., W. Gemert, S. K. Kangethe, and R. Juma. 1986. "Seasonal Children with Diarrhea in South Africa." Journal of Health, Population, Pattern of Some Causative Agents of Childhood Diarrhoea in Nairobi." and Nutrition 19: 199­203. East African Medical Journal 63: 373­81. Shamebo, D., L. Muhe, A. Sandstrom, and S. Wall. 1991. "The Butajira Mutanda, L. N., S. K. Kangethe, R. Juma, E. O. Lichenga, and C. Gathecha. Rural Health Project in Ethiopia: Mortality Pattern of the Under Fives." 1985. "Aetiology of Diarrhoea in Malnourished Children at Kenyatta Journal of Tropical Pediatrics 37: 254­61. National Hospital." East African Medical Journal 62: 835­42. Snow, R. W., J. R. Armstrong, D. Forster, M. T. Winstanley, V. M. Marsh, Nakano, T., F. N. Binka, E. A. Afari, D. Agbodaze, M. E. Aryeetey, J. A. A. C. R. Newton, C. Waruiru, I. Mwangi, P. A. Winstanley, and K. Marsh. Mingle, H. Kamiya, and M. Sakurai. 1990."Survey of Enteropathogenic 1992. "Childhood Deaths in Africa: Uses and Limitations of Verbal Agents in Children with and without Diarrhoea in Ghana." Journal of Autopsies." Lancet 340: 351­55. Tropical Medicine and Hygiene 93: 408­12. Snow, R. W., I. Bast de Azevedo, D. Forster, S. Mwankuyse, G. Bomu, Nakata, S., Z. Gatheru, S. Ukae, N. Adachi, N. Kobayashi, S. Honma, G. Kassiga, C. Nyamawi, T. Teuscher, and K. Marsh. 1993. "Maternal J. Muli, J. Nyangao, E. Kiplagat, P. M. Tukei, and S. Chiba. 1999. Recall of Symptoms Associated with Childhood Deaths in Rural East "Epidemiological Study of the G Serotype Distribution of Group A Africa." International Journal of Epidemiology 22: 677­83. Rotavirus in Kenya from 1991 to 1994." Journal of Medical Virology 58: Snyder, J. D., and M. H. Merson. 1982. "The Magnitude of the Global 296­303. Problem of Acute Diarrheal Disease: A Review of Active Surveillance Obi, C. L., A. O. Coker, J. Epoke, and R. N. Ndip. 1997. "Enteric Bacterial Data." Bulletin of the World Health Organization 60: 605­13. Pathogens in Stools of Residents of Urban and Rural Regions in Steele, A. D., H. R. Basetse, N. R. Blacklow, and J. E. Herrmann. 1998. Nigeria: A Comparison of Patients with and without Diarrhoea and "Astrovirus Infection in South Africa: A Pilot Study." Annals of Tropical Controls without Diarrhoea." Journal of Diarrhoeal Diseases Research Paediatrics 18: 315­19. 15: 241­47. Steele, A. D., A. Geyer, J. J. Alexander, H. H. Crewe-Brown, and P. J. Fripp. Okeke, I. N., A. Lamikanra, H. Steinrück, and J. K. Kaper. 2000. 1988. "Enteropathogens Isolated from Children with Gastro-Enteritis "Characterization of Escherichia Coli Strains from Cases of Childhood at Ga-Rankuwa Hospital, South Africa." Annals of Tropical Paediatrics Diarrhea in Provincial Southwestern Nigeria." Journal of Clinical 8: 262­67. Microbiology 38: 7­12. Sundary, R. 1986. "Health Implications of Sex Discrimination in Omondi-Odhiambo, J., K. Ginneken, and A. M. Voorhoeve. 1990. Childhood: A Review Paper and an Annotated Bibliography." "Mortality by Cause of Death in a Rural Area of Machakos District, WHO/UNICEF/FHE 86.2, WHO, Geneva. Kenya in 1975­78." Journal of Biosocial Science 22: 63­75. Tswana, S. A., P. H. Jorgensen, R. W. Halliwell, R. Kapaata, and S. R. Moyo. Oni, G. A., D. A. Schumann, and E. A. Oke. 1991. "Diarrheal Disease 1990. "The Incidence of Rotavirus Infection in Children from Two Morbidity, Risk Factors and Treatments in a Low Socioeconomic Area Selected Study Areas in Zimbabwe." Central African Journal of Medicine of Ilorin, Kwara State, Nigeria." Journal of Diarrhoeal Diseases Research 36: 241­46. 9: 250­57. United Nations. 2000. United Nations Millennium Declaration. General Paquet, C., P. Leborgne, A. Sasse, and F. Varaine. 1995. "An Outbreak of Assembly resolution 55/2, United Nations, New York. http://www. Shigella Dysenteriae Type 1 Dysentery in a Refugee Camp in Rwanda." un.org/millennium/declaration/ares552e.pdf. Santé 5: 181­84. Utsalo, S. J., F. O. Eko, and O. E. Antia-Obong. 1991. "Cholera and Vibrio Pennap, G., C. T. Pager, I. Peenze, M. C. de Beer, J. K. P. Kwaga, W. N. parahaemolyticus Diarrhoea Endemicity in Calabar, Nigeria." West Ogalla, J. U. Umoh, and A. D. Steele. 2002."Epidemiology of Astrovirus African Journal of Medicine 10: 175­80. Infection in Zaria, Nigeria." Journal of Tropical Pediatrics 48: 98­101. Vaahtera, M., T. Kulmala, K. Maleta, T. Cullinan, M.-L. Salin, and P. Pennap, G., I. Peenze, M. de Beer, C. T. Pager, J. K. P. Kwaga, W. N. Ogalla, Ashorn. 2000. "Epidemiology and Predictors of Infant Morbidity in J. U. Umoh, and D. Steele. 2000. "VP6 Subgroup and VP7 Serotype of Rural Malawi." Paediatric and Perinatal Epidemiology 14: 363­71. Human Rotavirus in Zaria, Northern Nigeria." Journal of Tropical Vesikari, T., D. O. Matson, P. Dennehy, P. Van Damme, M. Santosham, Pediatrics 46: 344­47. Z. Rodriguez, M. J. Dallas, et al. 2006. "Safety and Efficacy of a Perch, M., M. S. Sodemann Jakobsen, P. Vallentiner-Branth, H. Steinland, Pentavalent Human-Bovine (WC3) Reassortant Rotavirus Vaccine." T. K. Fisher, L. D. Duarte, P. Aaby, and K. Mølbak. 2001. "Seven Years' New England Journal of Medicine 354: 23­33. Experience with Cryptosporidium parvum in Guinea-Bissau, West Victora, C. G., J. Bryce, O. Fontaine, and R. Monasch. 2000. "Reducing Africa." Annals of Tropical Paediatrics 21: 313­18. Deaths from Diarrhea Through Oral Rehydration Therapy." Bulletin of Pison, G., J. F. Trape, M. Lefebvre, and C. Enel. 1993 "Rapid Decline in the World Health Organization 78: 1246­55. Child Mortality in a Rural Area of Senegal." International Journal of Victora, C. G., S. R. Huttly, S. C. Fuchs, F. C. Barros, M. Garenne, O. Leroy, Epidemiology 22 (1): 72­80. O. Fontaine, J. P. Beau, V. Fauveau, and H. R. Chowdury. 1993. Rowland, S. G. J., N. Lloyd-Evans, K. Williams, and M. G. M. Rowland. "International Differences in Clinical Patterns of Diarrheal Deaths: A 1985. "The Etiology of Diarrhea Studied in the Community in Young Comparison of Children from Brazil, Senegal, Bangladesh, and India." Urban Gambian Children." Journal of Diarrhoeal Diseases Research 3: Journal of Diarrhoeal Diseases Research 11 (1): 25­29. 7­13. Victora, C. G., M. T. Olinto, F. C. Barros, and L. C. Nobre. 1996. "Falling Ruiz-Palacios, G. M., I. Perez-Schael, R. Velazquez, H. Abate, T. Breuer, Diarrhea Mortality in Northeastern Brazil: Did ORT Play a Role?" S. A. Costa-Clemens, B. Cheuvart, et al. 2006. "Safety and Efficacy of an Health Policy and Planning 11: 132­41. 122 | Cynthia Boschi-Pinto, Claudio F. Lanata, Walter Mendoza, and Demissie Habte Victora, C. G., A. Wagstaff, J. A. Schellenberg, D. Gwatkin, M. Claeson, ------. 2002. The World Health Report 2002: Reducing Risks, Promoting and J.-P. Habicht. 2003. "Applying an Equity Lens to Child Health Healthy Life. Geneva, WHO. Mortality: More of the Same Is Not Enough." Lancet 362: 233­41. ------. 2005. "Mothers and Children Matter--So Does Their Health." In Walker, N., B. Schwartländer, and J. Bryce. 2002. "Meeting International The World Health Report 2005--Make Every Mother and Child Count. Goals in Child Survival and HIV/AIDS." Lancet 360: 284­89. Geneva: WHO. Watts, T., N. Ng'andu, and J. Wray. 1990. "Children in an Urban Township Yeager, B. A., S. R. Huttly, R. Bartolini, M. Rojas, and C. F. Lanata. 1999. in Zambia. A Prospective Study of Children during Their First Year of "Defecation Practices of Young Children in a Peruvian Shanty Town." Life." Journal of Tropical Pediatrics 36: 287­93. Social Science and Medicine 49: 531­41. WHO (World Health Organization). 2000. WHO-UNICEF-WSSCC Global Water Supply and Sanitation Assessment 2000 Report. Geneva: WHO. Diarrheal Diseases | 123 Chapter 10 Developmental Disabilities Geoff Solarsh and Karen J. Hofman Developmental Disabilities are disorders of the developing MEDICAL AND SOCIAL MODELS OF DISABILITY nervous system that manifest during infancy or childhood as developmental delay or as limitations of function in one The definition just given has its origins in the medical model or multiple domains, including cognition, motor perform- of disability, which views disability as a problem of the per- ance, vision, hearing and speech, and behavior. Because of son, directly caused by disease, trauma, or other health con- the variable nature, extent, and timing of the disorders dition and requiring individual medical care from health or in the developing nervous system their clinical expression rehabilitation professionals. Management of the disability is varies enormously from one individual to another, both in aimed at cure or, more realistically, at producing personal severity and in relative effect on the different areas of func- adjustment or behavior change by the individual in tion. Because developmental disabilities are a composite of response to the disability. In contrast, the social model sees a large number of different health conditions, primary and the issue mainly as a socially created problem attributable to secondary prevention strategies vary for each of the compo- environmental and contextual factors, such as lack of aware- nent conditions, whereas tertiary prevention strategies, ness or social stigma in the broader society and deficient which address later effects on capacities in broad areas of social policies and legislation. These factors together create function, are largely shared across disorders. These disabili- an environment for people with disabilities that limits activ- ties are likely to continue indefinitely and to result in sub- ity and restricts participation. In the medical model inter- stantial limitations on many life activities, such as affected vention usually means the prevention and early treatment of individuals' ability to care for themselves, express and re- health conditions in the individual (primary and secondary ceive language, learn, be mobile, and live independent and prevention), whereas in the social model it means the pro- economically self-sufficient lives. motion of functional capacity and the achievement of full 125 Figure 10.1 Relationship between Impairment, Disability, Figure 10.2 Relationship between Body Functions, Activities, and Handicap (ICIDH) and Participation (ICF) Disease or Health condition Impairment Disability Handicap Injury (disorder or disease) Source: WHO 1980. participation in the physical and social environment (terti- Body ary prevention). In this chapter an attempt is made to inte- functions and Activities Participation grate these perspectives. structures The International Classification of Functioning In 1980 the World Health Organization (WHO) published Environmental Personal factors factors the International Classification of Impairments, Disabilities and Handicaps (WHO 1980) as a companion classification Source: Authors. to the International Classification of Diseases to document the consequences of congenital or acquired illnesses with outcomes of children and young adults. The neurological nonfatal outcomes (figure 10.1). These consequences were and developmental deficits that have their primary origin in evident in impairments, defined as abnormalities of body adverse social and environmental conditions, such as or organ structures and functions; disabilities, defined as poverty, poor nutrition, and social deprivation, during the reduction of a person's ability to perform basic and everyday critical years of early brain growth and development usually tasks; and handicaps, defined as a person's disadvantage in present later in childhood as cognitive impairments and fulfilling social roles. This framework suggested a hierarchi- poor performance at school. This group of at-risk infants cal and linear relation between these three dimensions that and children is likely to exceed by many times the number entailed a problem-solving sequence in which interventions of children with readily identifiable developmental disabili- at one level have the potential to modify succeeding ele- ties. Because the contribution of this group to the burden of ments (Fryers 1992). developmental disabilities is difficult to measure and the Two of the most important criticisms of this classifica- interventions are more closely linked to broad community tion that are particularly relevant for this review are its fail- development and poverty alleviation strategies, they will not ure to adequately cover disabilities affecting children and its be considered, except in passing, in this chapter. This in no limited utility for public health applications. A closely relat- way implies that they have any less a priority in policies and ed criticism has been its failure to take into account the per- programs that address developmental disabilities in low- vasive role of the environment in exacerbating or reducing income countries. the nature and extent of disablement. In its revised International Classification of Functioning, THE CHANGING BURDEN OF DISEASE Disability and Health the WHO (2001) has sought to IN CHILDHOOD address some of these criticisms and, in particular, to unify the apparent polarities in the medical and social models of Developmental disabilities and their prevention have not disability (figure 10.2). In the International Classification of had much prominence on the public health agenda in devel- Functioning a person's functioning and disability is con- oping countries over the past three or four decades. Steady ceived as a dynamic interaction between health conditions shifts in the patterns of mortality and morbidity over this (diseases, disorders, injuries, and trauma) and the effect of period are now beginning to challenge these traditional contextual and environmental factors in limiting activity public health priorities. and restricting participation. The Health Transition The Effect of Poverty Infant and child mortality has steadily declined in Sub- In developing countries serious developmental disabilities Saharan Africa countries during the second half of the twen- represent only a proportion of the poor developmental tieth century (Ahmad, Lopez, and Inoue 2000; Hill et al. 126 | Geoff Solarsh and Karen J. Hofman 1999) along with an accompanying fall in fertility (Cleland, cases, the increase is attributed to the human immunodefi- Onuoha, and Timaeus 1994; Cohen 1993). This is a result of ciency virus (HIV) epidemic (Nicoll et al. 1994; United a complex process of social and economic change in these Nations Population Division 1999). Recent evidence sug- countries coupled with sustained implementation of family gests that the proportion of under-five mortality attributa- planning programs (United Nations Population Division ble to HIV and the acquired immune deficiency syndrome 1998) and primary health care interventions (Fauveau et al. (AIDS) in Sub-Saharan Africa is quite variable (0.1 to 1990; Grant 1992; Velema et al. 1991) that specifically target 42.4 percent) and that it is highest in some of the countries preventable diseases in infancy and childhood. with the steepest reductions in postneonatal mortality in the This "demographic and epidemiological transition" has past two decades (Walker, Schwartlander, and Bryce 2002). occurred at varying rates in different African countries. These data suggest that programs to tackle developmen- Infant mortality trends during the health transition in tal disabilities may have a greater claim on national resources developed countries have shown large absolute decreases in in some Sub-Saharan Africa countries than in others during postneonatal mortality accompanied by increases in the this complex health transition. But even in countries with proportional contribution of neonatal mortality to overall high under-five mortality and HIV seroprevalence rates, 8 or infant mortality (MacFarlane and Mugford 1984). These 9 out of every 10 children will survive beyond five years. shifts in the ratio of neonatal to postneonatal mortality Many of them will continue to be at risk for developmental have also occurred in countries in Sub-Saharan Africa with disabilities because of preventable biological factors and the the lowest infant mortality rates. Neonatal mortality now lack of services and programs to identify, treat, and amelio- accounts for 45 percent of all infant deaths in South Africa rate the impact of these disabilities. Additionally, as will be (SADHS 1998, 7). These demographic changes suggest that discussed later, children with HIV infection may be at special a higher priority should be given to programs that address risk for developmental disabilities. perinatal conditions. They also signify a need to move away from an exclusive preoccupation with child survival strate- Measuring the Burden of Developmental Disabilities gies to those that aim to improve the quality of survival in the 92 to 95 percent of children now surviving beyond five Developmental disabilities, because of their early onset and years in many African countries. Pointing to pregnancy and lifelong requirement for support and care, impose enor- the neonatal period targets high-risk groups not only for mous social and economic burdens on affected individuals, residual infant mortality but also for developmental disabil- their families, and their communities. Calculation of the loss ities, because many of the most important causes of these of disability-adjusted life years (DALYs) is now the method disabilities are addressed by interventions during this widely used to capture and compare the combined effect on period. the global burden of disease (GBD) of premature mortality and decreased functional capacity resulting from designated health conditions (Murray 1994). An important obstacle to The Impact of HIV the measurement of the burden of developmental disabili- Arguments for the assignment of a higher priority to ties in low-income countries is the lack of good quality research and programs that address developmental disabili- prevalence data. An additional concern stems from the char- ties in low-income countries are regularly based on evidence acteristics and applications of the measurement itself. for the health transition in these countries. These transi- In burden-of-disease calculations the years of life lost tions, however, are generally uneven both between and for each death are estimated and assigned a relative value within countries, and this is particularly true of the African based on the age at death (age weighting). Age weighting is subcontinent. The "unfinished agenda" of infectious dis- intended to reflect differential productivity of an individual eases and malnutrition continues to make a variable contri- at different stages of his or her life cycle and, thereby, assigns bution to postneonatal mortality and morbidity in the dif- a different social value to lives lived at different ages. A year ferent countries and their subpopulations, depending on lived at age 2 counts for only 20 percent of a year lived at 25, social, economic, and political conditions and the coverage when the age-weighting function is at a maximum. The effect and quality of primary health care services. Many African of this is to reduce the DALYs lost by premature death of chil- countries are now reporting rising postneonatal mortality dren with developmental disabilities and to potentially lower with patterns of disease often indistinguishable from those the relative importance of these conditions in decisions in the previous decade (Ahmad et al. 2000), and in many about resource allocations (Anand and Hanson 1997). Developmental Disabilities | 127 To measure the additional contribution of disability, the sample of those screening negative undergo a detailed number of years of healthy life lost is estimated by multiply- medical and psychological examination from which an eti- ing the expected duration of the condition (to remission or ological diagnosis is made and disability rated (none, mild, death) by a disability weighting (0­1). Disability weighting moderate, and severe) in each of the following areas: gross does not take into account the way that individual and social motor, fine motor, vision, hearing, cognition, speech, and resources can compensate for the level of disability experi- seizures. enced. The failure to factor in these compensatory mecha- Two-phase surveys of this kind have been shown to have nisms distorts burden-of-death estimates and undervalues a high sensitivity but a low positive predictive value. They the therapeutic and rehabilitation benefits offered by reha- are, therefore, ideally suited to serve as screening instru- bilitation professionals and programs (Jelsma, De Weerdt, ments but completely inappropriate as case-finding tools for and De Kock 2002). epidemiological studies or as the basis for referring children Developmental disabilities incorporate many different with disabilities for rehabilitation services. Using the ten- disease entities with their different causes; time of onset; question screen as the only basis for determining prevalence natural histories; amenability to primary, secondary, and of disabilities has been shown to overestimate prevalence by tertiary prevention; and the relative risk of premature mor- up to 300 percent (Durkin, Hasan, and Hasan 1995). tality and functional limitation. Since the burden-of-disease Many clinical researchers have set the ethical requirement methodology is based on calculations for individual disease that all children screening positive in a target population entities, each of these would need to be addressed separately should be included in the second phase so that they can be in order to arrive at a burden-of-disease estimate for devel- referred to medical and rehabilitation services. Although no opmental disabilities as a whole. Without this information, similar imperative applies to those screening negative, the more pressing and immediate issues, such as HIV/AIDS and inclusion of negatives is necessary (thus the minimal sample malaria, will dominate the health agenda and displace of 10 percent) because, using this method, the prevalence developmental disabilities from consideration for resource estimate for the total population is obtained as a weighted allocation. average of the rates in those screening positive versus those screening negative (Shrout and Newman 1989). PREVALENCE AND INCIDENCE Disability Prevalence Data in Sub-Saharan Africa In low-income countries estimates of the frequency of devel- Reliable data based on the criteria set above on the preva- opmental disabilities invariably come from cross-sectional lence of developmental disabilities is scarce in the African surveys, which measure prevalence, or the number of existing subcontinent. An initial Medline search was combined with cases in the population. Although data on incidence, or the a search of two major specialist disability journals, Disability number of newly occurring cases, provide a better measure of and Rehabilitation and The International Journal of true frequency in populations, repeat assessments of the Rehabilitation Research, for the years 1995 to 2004 and a same children over shorter periods of time are not cost- detailed review of all peer-reviewed publications within effective in most low-income countries. which the two-phase methodology is employed to measure the prevalence of developmental disabilities in developing countries. In the relatively few studies identified in Sub- Two-Phase Childhood Disability Surveys Saharan Africa, not a single one fulfilled the methodological The best available standard for the measurement of the criteria laid out above for two-phase disability studies. prevalence of developmental disabilities in population- Most of the identified studies were conducted in South based surveys is the two-phase survey method, which has Africa. The prevalence rates for all categories of develop- been validated in many different population settings in the mental disability in these studies varied from 11 to 60 per developing world (Durkin et al. 1994; Thorburn, Desai, 1,000 children (Cornielje et al. 1993; Couper 2002; Irlam and Davidson 1992; Zaman et al. 1990). The first phase 1996; Katzenellenbogen et al. 1995; Schneider 1997). consists of a survey of all households in a drawn sample or Although two-phase surveys were used in most studies, the target population in which all children are screened using phase 1 screening questionnaires were not standardized in the "ten questions" questionnaire. In the second phase all cases and were not validated for use in African popula- all children screening positive and a random 10 percent tions in any of the studies (table 10.1). 128 | Geoff Solarsh and Karen J. Hofman Table 10.1 Prevalence of Disability in Sub-Saharan Africa Overall Speech, Year Country Author(s) Methods prevalence Cognition Motor Vision hearing Behavior Epilepsy (per 1,000 target population) 1987 Zambia Stein, 2-phase survey -- 35.0 (total) -- -- -- -- -- Belmont, and TQ questionnaire 30.0 (mild) Durkin 1987 Medical exam 5.0 (severe) 3­9 years 1992 South Cornielje et al. 2-phase survey 11.0 -- -- -- -- -- -- Africa, 1993 No medical exam (no severity Gelukspan 5­37 months rating) 1995 South Katzenellen- 2-phase survey 44.0 -- -- -- -- -- -- Africa, bogen et al. No medical exam (moderate to Western 1995 All ages severe) Cape 1995 South Africa Irlam 1996 2-phase survey 16.9 -- -- -- -- -- -- KwaZulu TQ questionnaire (no severity No medical exam rating) 2­19 years 1997 South Africa Schneider 1-phase survey 16.0 (1­5 years) -- -- -- -- -- -- 1997 Questionnaire 32.0 (6­10 years) All ages 45.0 (11­15 years) 2000 South Africa Christianson 2-phase survey -- 35.5 (total) -- -- -- -- 7.3 Bushbuckridge et al. 2002 TQ questionnaire 29.1 (mild) Medical exam 6.4 (severe) 2­9 years 2002 South Africa Couper 2002 2-phase survey 60.0 -- 28.0 2.0 24.0 (speech) 37 4.0 KwaZulu TQ questionnaire 20.0 (hearing) No medical exam 0­10 years Source: Authors. Note: TQ ten questions; -- not available. A more serious limitation is the variation in the methods A tabulation of some of the two-phase studies cond- or content of the evaluation in the second phase. These clin- ucted elsewhere in the developing world is provided as ical assessments were not standardized, and they varied table 10.2 for comparison and similarly shows wide varia- from repeat questionnaires by fieldworkers to assessments tions in prevalence rates for all categories of disability in by rehabilitation assistants, rehabilitation professionals, and childhood from different studies (Durkin et al. 1994; neurodevelopmental pediatricians. Additionally, for valid Durkin, Hasan, and Hasan 1995; Islam, Durkin, and calculations of prevalence in two-phase studies a 10 percent Zaman 1993; Milaat et al. 2001; Natale et al. 1992; Tamrat sample of individuals screening negative should be included et al. 2001; Zaman et al. 1990). The studies in Bangladesh, in the second phase. This did not occur in any of the cited Jamaica, and Pakistan that provided the range in preva- studies, nor was it always clear from the description of the lence of from 8.1 to 31.0 per 1,000 children were based on methods how prevalence rates were derived. A final concern identical and established methodologies, used the same set is the lack of a standardized approach in most studies to the of instruments, and relied on identical rating systems for grading of severity of disability. Since it is clear from care- severity. These rates in all likelihood reflect true differ- fully conducted studies elsewhere that the ratio between ences in the prevalence of developmental disabilities in severe and mild disabilities may be very high, comparisons these three countries and provide a strong platform on between studies that include children of differing severity which to begin to look at types, determinants, and causes may lead to erroneous conclusions about the true size and of disability and on which to build a rational approach to nature of the problem in different areas. intervention. Developmental Disabilities | 129 Table 10.2 Prevalence of Disability in Other Developing Countries Overall Speech, Year Country Author(s) Methods prevalence Cognition Motor Vision hearing Learning Epilepsy (per 1,000 target population) 1990 Bangladesh Zaman et al. 2-phase survey 16.0 -- -- -- -- -- -- 1990 TQ questionnaire (moderate to Medical exam severe) 2­9 years 1992 India Natale et al. 1-phase survey 172.0 (poorer) -- -- -- -- -- -- 1992 TQ questionnaire 82 (richer) 2­9 years 1993 Bangladesh Islam, Durkin, 2-phase survey -- 5.9 (severe) -- -- -- -- -- and Zaman TQ questionnaire 1993 Medical exam 2­9 years 1994 Bangladesh 8.1 -- -- -- -- -- -- (moderate to severe) Jamaica Durkin et al. 2-phase survey 19.8 -- -- -- -- -- -- 1994 TQ questionnaire (moderate to Medical exam severe) Pakistan 2­9 years 31.0 -- -- -- -- -- -- (moderate to severe) 1995 Pakistan Durkin, Hasan, 2-phase survey 44.3 19.0 (severe) 19.5 15.0 5.2 (hearing) -- 5.0 and Hasan TQ questionnaire 1995 Medical exam 2­9 years 2001 Saudi Arab Milaat et al. 1-phase survey 36.7 -- -- -- -- -- -- 2001 TQ questionnaire (no severity 0­15 years rating) 2001 Ethiopia Tamrat et al. 1-phase survey 31 -- -- -- -- -- -- 2001 TQ questionnaire (no severity 5­14 years rating) Source: Authors. Note: TQ ten questions; -- not available. Cognitive Disabilities disability surveys come from Zambia (Stein, Belmont, and Population-based prevalence data on cognitive disabilities Durkin 1987), where rates of 5 per 1,000 were recorded for are sparse in developing countries, and what little is avail- severe MR and 30 per 1,000 for mild MR, and in a recent able comes mainly from outside Sub-Saharan Africa. Severe study from rural South Africa (Christianson et al. 2002), mental retardation (MR), defined as a decreased general where similar rates of 6.4 per 1,000 were noted for severe intelligence quotient of less than or equal to 55, accompa- MR and 29.1 per 1,000 for mild MR. These figures are a little nied by significant limitations in adaptive capability, is con- lower than suggested averages for severe MR in developing sistently found to be in the range of 3 to 5 per 1,000 persons countries (9.3 per 1,000) but similar to suggested averages in developed countries. In the few available studies from for mild MR (29.8 per 1,000; Roeleveld, Zielhuis, and low-income countries, rates are significantly higher and Gabreels 1997). The failure to ascertain a specific biological range from 6 per 1,000 and 22 per 1,000 for severe MR and cause in many of these children suggests that many cognitive 14.5 per 1,000 and 65.3 per 1,000 for mild MR in disabilities may have their origins in maternal and infant Bangladesh (Islam, Durkin, and Zaman 1993) and Pakistan malnutrition and impoverished environments, which have (Durkin, Hasan, and Hasan 1998), respectively. The only pervasive adverse effects on growth and psychological equivalent figures from well-designed population-based development. 130 | Geoff Solarsh and Karen J. Hofman Motor Disabilities blind supports this view (O'Sullivan, Gilbert, and Foster 1997). An overall estimate of blindness prevalence was 0.35 No methodologically sound studies could be found that to 0.6 per 1,000 children. Although 39 percent of causes of reported population-based prevalence estimates for motor blindness or severe visual impairment were found to be pre- disabilities or its subtypes in Sub-Saharan Africa. One ventable, only 5 percent of the affected children had condi- report, which lacked a detailed second-phase medical tions amenable to primary preventive measures, such as evaluation and did not include children screening negative, vitamin A deficiency or measles. Almost a quarter of the set the prevalence at 28 per 1,000 in a rural South African children (23 percent) had inherited conditions, intrauterine district (Couper 2002). A comparative rate of 19.5 per 1,000 infections, or retinopathy of prematurity; the majority of for severe motor disabilities and 52.5 per 1,000 for mild these problems were potentially preventable through genetic motor disabilities was reported from Pakistan. There is a counseling and improved antenatal and neonatal care. clear and immediate need for studies to document inde- The remaining 11 percent needed sophisticated surgery pendent prevalence rates for motor disabilities, given its for such conditions as cataracts of unknown origin and common association with severe cognitive disabilities and glaucoma. the observation from studies elsewhere that it is an important and common functional limitation in children with developmental disabilities (Durkin, Hasan, and Hasan 1995). Hearing Disabilities Estimates of hearing loss and profound deafness in devel- oped countries are on the order of 1 per 1,000, compared Vision Disabilities with 1.4 to 4.0 in developing countries. The WHO estimates There are estimated to be 1.5 million blind children world- that there are 120 million people worldwide with hearing wide (WHO 1992). The prevalence of blindness in children impairment, and 78 million of those affected are in devel- in European countries varies from 0.2 to 0.4 per 1,000 chil- oping countries (WHO 1995b). As in other domains, there dren and in African countries from 0.5 to 1.1 per 1,000 are few prevalence data on hearing disabilities in Sub- (Gilbert et al. 1999). Several studies have estimated that as Saharan Africa. Where data are available, it is often difficult much as 47 percent of blindness or severe visual impairment to separate methodological limitations from true differ- in developing countries is preventable or curable (Adeoye ences in prevalence between these studies (WHO 1995b). 1996; Nwosu 1998; Silver et al. 1995). Data from studies Reported rates of profound hearing loss range from 2.1 per done in schools for the blind in East, central, and West 1,000 (sensorineural loss) in Swaziland (Swart et al. 1995) African countries show that the most common causes of to 4.0 per 1,000 in Sierra Leone (Seely et al. 1995; see blindness in children are acquired diseases, such as vitamin table 10.3). Chronic otitis media has been demonstrated to A deficiency (29 percent) and measles (27 percent) (Gilbert be the most frequent cause of hearing impairment in many et al. 1993; Gilbert et al. 1995). developing countries (Smith et al. 1996). Estimates suggest Quite variable rates might be expected in countries at dif- that as much as 50 to 66 percent of all hearing impairment ferent stages in the health transition. A recent study on the in the developing world is preventable (Smith and Hatcher prevalence of blindness in South African schools for the 1992). Table 10.3 Prevalence of Hearing Disability in Sub-Saharan Africa Hearing disability Year Country Author(s) Methods (per 1,000 target population) 1985 The Gambia McPherson and Holborow 1985 Schoolchildren 2.7 (severe to profound) 1987 Tanzania Manni and Lema 1987 Schoolchildren 3.5 (severe to profound) 1995 Sierra Leone Seely et al. 1995 Population-based survey (5­15 years) 4.0 (profound) 1995 Kenya Hatcher et al. 1995 Schoolchildren 2.4 (profound) 1995 Swaziland Swart et al. 1995 Schoolchildren (class 1) 2.1 (sensorineural loss) Source: Authors. Developmental Disabilities | 131 Learning Disabilities period of rapid brain growth and development during pregnancy, infancy, and childhood. They may have their A learning disability is traditionally defined as a disorder in early beginnings in the genetic makeup of the parents, in the one or more of the basic psychological processes involved in nutritional status of the mother throughout her life cycle, in understanding or in using language, spoken or written, maternal health conditions and environmental exposure resulting in an imperfect ability to listen, think, speak, read, during pregnancy, or in an early or abnormal birth process. write, spell, or do mathematical calculations. The definition They may be the consequence of adaptation difficulties soon specifically excludes learning problems that are primarily after birth; infections, poor nutrition, and injuries in infancy the result of visual, hearing, or motor disabilities and those and childhood; or the complex and pervasive effects of resulting from mental retardation or emotional disturbance. adverse social and environmental conditions in impover- In Sub-Saharan Africa it may be difficult to distinguish ished communities (figure 10.3). These factors may act between traditional learning disabilities and the conse- singly or in combination and may make contributions of quences of adverse social and environmental conditions. varying magnitude, depending on background infant and No studies could be found that described the prevalence under-five mortality rates, the quality of health services, and of learning disabilities in Sub-Saharan Africa. A Cape Town special risk factors that may apply in individual locations study in a population not typical of most of the rest of the or subpopulations. These factors may also vary in their subcontinent found that the origin of 45 percent of learning amenability to intervention. disabilities was prenatal, 17 percent was perinatal, 9 percent An understanding of these factors and their population- was postnatal, and about 25 percent was unknown (Molteno attributable risk is a critical preamble to the development of and Lachman 1996). The distribution between these etio- strategies for primary and secondary prevention. This sec- logical categories varied by ethnic group; children from tion highlights risk factors for developmental disabilities for white and mixed race families had relatively high prenatal which evidence exists from Sub-Saharan Africa or other contributions (55 percent) compared with children from developing countries of high prevalence or public health black African families (23 percent). These relative contribu- and economic impacts, or both, and viable, if unrealized, tions were reversed in the perinatal category; black African potential for prevention. children contributed 37 percent, and children from mixed race groups, 8 percent. Similar distributions were found in an earlier Zimbabwean study but with a much higher pro- Congenital Disorders portion being of unknown cause (Axton and Levy 1974). Congenital disorders are defined here as any potentially dis- Forty percent of all cases were considered to be preventable. abling condition arising before birth and including those Many of the children in these studies had other major dis- caused by environmental, genetic, and unknown factors, abilities, including motor, cognitive, and sensory deficits, whether they are evident at birth or become manifest later raising doubts as to whether they fit the traditional defini- in life. It does not include congenital infections or nutri- tion for this disorder. tional factors influencing intrauterine growth, which will be As developing countries go through the health transition, discussed later. it seems likely that learning disabilities will become an The frequency of congenital disorders is best described in increasingly important concern as the countries begin to terms of birth prevalence, that is, affected births per 1,000 in make more qualitative investments in the future human the absence of a prevention program. Since there are no potential and productivity of their populations. In a coun- prevalence data from developing countries, projections are try like South Africa, where economic development and based on data from developed countries; adjustments rapid political change have coincided, the great demand for are made for regional differences in the prevalence of access to educational opportunities has highlighted the hemoglobin disorders (WHO 1994), glucose-6-phosphate learning deficits and, in some cases, the disabilities of many dehydrogenase deficiency (Luzzato and Mehta 1989), and children. the effect of customary consanguineous marriage (Alwan and Modell 1997), all of which have their greatest burden in DETERMINANTS AND RISK FACTORS less-developed parts of the world (table 10.4). The estimated rate of 61 per 1,000 in Sub-Saharan Africa, which approxi- Developmental disabilities have a wide range of origins, mates or exceeds infant mortality in some Sub-Saharan occurring from the time of conception through an extended African countries, is inflated by the inclusion of fetal losses, 132 | Geoff Solarsh and Karen J. Hofman Figure 10.3 Research Steps in the Development of Public Health Interventions Conception Nutritional Alcohol Nutrition supplementation Genetics Infections Treatment of infections Drugs Hypertension Antenatal screening and obstetric care Smoking Preeclamptic toxaemia Delivery Low Congenital birthweight abnormali- Meningitis ties Head injury Perinatal Congenital asphyxia infections Diarrhea HIV/AIDS Neonatal screening Pneumonia Growth monitoring and nutrition education Undernutrition Developmental screening Poverty Deworming and micronutrient supplementation Poor maternal Maternal education education and support Early child Inadequate stimulation and stimulation education Social welfare support Abnormal INTERVENTIONS development and RISK FACTORS disability Source: Authors. Table 10.4 Estimated Birth Prevalence of Infants with Serious Congenital Disorders, by WHO Region Births per Population year Congenital Chromosomal Single gene Total congenital Annual (millions) (millions) malformations disorders per disorders disorders affected live WHO region 1996 1996 per 1,000 1,000 per 1,000 per 1,000 births Eastern Mediterranean 506 18.1 35.7 4.3 27.3 69.0 1,237,225 African 540 23.0 30.8 4.4 25.0 61.0 1,412,427 South East Asian 1,401 38.2 31.0 3.9 14.7 51.0 1,946,606 European 867 10.8 31.3 3.7 12.4 49.0 522,832 American 782 16.2 30.9 3.8 11.9 48.0 774,235 Western Pacific 1,650 31.3 30.6 3.5 11.4 47.0 1,464,067 Total 5,746 137.6 31.5 3.9 16.8 53.0 7,357,392 Source: WHO 1999. Developmental Disabilities | 133 but it does provide some indication of the increasing Africa is likely to be on the order of 25 per 1,000 (table 10.4). contribution of congenital disorders to burden of disease The collective impact is significant, but none of the individ- and disability as infant mortality rates fall (WHO 1999). ual disorders is a public health problem with sufficiently In West Africa 2 to 3 percent of all children have a serious high population-attributable risk to currently merit target- hemoglobinopathy (sickle-cell anemia, thalassemia) ed intervention. (Adeoye 1973; Obama et al. 1994). These children are at risk for nervous system complications, the frequency of which Perinatal and Neonatal Conditions may be as high as 12.8 percent. Complications of sickle-cell disease in children include mental changes, cerebrovascular Perinatal events, such as preterm delivery, low birthweight, accidents, cranial nerve palsies, dural sinus thrombosis, intrauterine growth retardation (IUGR), and birth asphyxia and increased susceptibility to meningitis, especially salmo- or injury, are commonly associated with an elevated risk of nella and pneumococcal meningitis. A recent study in the early neonatal death and, in those who survive, of impaired United States revealed that 33 percent of children observed physical, sensory, or mental development in infancy and with sickle-cell disease had mild mental retardation (Steen childhood. Many factors, acting singly or in combination, et al. 1999). contribute to the elevated frequency of these events in low- Down syndrome has until quite recently been regarded as income countries. These include the effect on the growing rare in black African populations. Reports of a birth preva- and developing fetus of maternal macro- and micronutrient lence of 1.16 (Adeyokunnu 1982) and 2.09 (Venter et al. deficiencies before and during pregnancy; the direct and 1995) per 1,000 from Nigeria and South Africa, respectively, indirect effects of maternal systemic and genital tract infec- suggest that it may be more common than previously tions, such as syphilis, rubella, cytomegalovirus, and malar- thought. These figures may also underrepresent the true ia; and the neurological effects of low blood glucose, hypox- prevalence, because evidence suggests that the syndrome ia, bilirubin toxicity, and acquired infections in the first few in some of these children may go unrecognized or unde- days of life. Many of these risks may be aggravated or ame- clared in rural black communities (Christianson and liorated, depending on the availability and quality of ante- Kromberg 1996). The prevalence is likely to be higher in natal, delivery, and postnatal services. Sub-Saharan Africa countries that have high fertility rates, In areas in which maternal and neonatal services are poor where effective family planning programs and prenatal and birth asphyxia is an important cause of developmental screening programs are often lacking and where 11 to disabilities, operational research to develop and evaluate 15 percent of births occur to mothers over the age of alternative approaches to the delivery of these services 35 years (Drugan et al. 1999). should be a first priority. In settings in which adequate The reported incidence of neural tube defects varies from maternal and neonatal services are available, research is country to country, from region to region within the same urgently needed on the etiology and prevention of adverse country, and from time to time (Windham and Edmonds perinatal outcomes, such as low birthweight, preterm birth, 1982). A reported prevalence for neural tube defects of 7 per and IUGR; on the causal pathways between these factors 1,000 children in Nigeria may have overestimated the inci- and developmental disabilities; and on the differential dence, because the study was based at a tertiary referral hos- impact of their prevention on the prevalence of neurodevel- pital (Windham and Edmonds 1982). The only available opmental disabilities in low-income countries. In undertak- population-based estimate of 3.35 per 1,000 comes from a ing this research it will be important to define IUGR and previously cited study in rural South Africa. its subtypes (Goldenberg et al. 1989); distinguish between Although consanguineous marriages are considered to be IUGR and its antecedents, many of which are independent an important cause of congenital abnormalities in many risk factors for poor neurodevelopmental outcome (Breart parts of the world, having particularly high prevalence in and Poisson-Salomon 1988); control for poor social and parts of South Asia, the Middle East, and North Africa, the environmental conditions that operate postnatally and that contribution of such marriages to the burden of develop- may modify neurodevelopmental outcomes (Breart and mental disabilities appears to be relatively small in most Poisson-Salomon 1988); select a set of outcomes that are parts of Sub-Saharan Africa. It has been extrapolated from sufficiently prevalent, well defined, and stable over time so the birth incidence of single gene disorders in developed that they can be measured with precision at defined time countries that the equivalent birth incidence in Sub-Saharan points; and use reasonable sample sizes. 134 | Geoff Solarsh and Karen J. Hofman INFECTION include deafness, interstitial keratitis, and mental retarda- tion. It is largely preventable through screening and ade- Numerous prenatal, perinatal, and postnatal infections can quate treatment in pregnancy of the 4 to 15 percent of damage the developing nervous system or sensory pathways women known to be affected in Sub-Saharan Africa (Schulz, and cause long-term disabilities in children and young Cates, and O'Mara 1987). adults. The relative contribution of these infections to the In spite of the availability of an established and highly burden of developmental disabilities is likely to vary from cost-effective intervention, it was found, in a recent survey country to country, influenced by overall infant mortality, of 22 countries in Sub-Saharan Africa, that only 38 percent postneonatal contribution to infant mortality, and regional of women attending antenatal clinics were being screened differences in the distribution of the infections known to be and treated for syphilis (Gloyd, Chai, and Mercer 2001). It associated with neurological sequelae during these different has been roughly estimated that every year up to 600,000 periods in the early life cycle. opportunities are missed to reduce adverse fetal and infant outcomes in Sub-Saharan Africa. Although reductions in Congenital Rubella fetal wastage and neonatal mortality may be the main bene- Congenital rubella is a major global cause of preventable fits, more effective antenatal treatment will also reduce hearing impairment, blindness, and intellectual disability. defined risks for developmental disabilities in these Mathematical modeling has yielded a global disease burden children. estimate for congenital rubella syndrome (CRS) of 110,000 Other congenital infections such as cytomegalovirus, tox- to 300,000 new cases per year (Cutts and Vynnycky 1999). oplasmosis, and herpes infections are also responsible for The incidence of CRS has been variably set at 0.5 to 2.2 per important neurological sequelae, but because they occur less 1,000 live births in developing countries during epidemics, frequently and are less amenable to primary or secondary which occur every four to seven years (Cutts et al. 1997). prevention, they are given less weight in this review. Although many developed countries have set elimination goals, only 28 percent of developing countries routinely vac- HIV Infection cinate against rubella (Robertson et al. 1997). No countries HIV infection is known to have an adverse effect on the in Sub-Saharan Africa include rubella in their national developing central nervous system and could potentially immunization program, and rubella serology, which is make a substantial contribution to the burden of develop- essential for surveillance, is unavailable in much of the sub- mental disabilities in populations with high HIV seropreva- continent (Robertson 2000). lence. HIV infection causes damage to the central nervous It is recommended that countries wishing to undertake system through direct cytopathic effects such as occurs prevention programs for CRS should either mount vaccina- in HIV-associated encephalopathy (Brustle et al. 1992), or tion programs for adolescent girls or women of reproduc- as a result of vasculopathy or immune-mediated factors. tive age or offer universal vaccination in infancy as part of Secondary complications of immunodeficiency, such as routine childhood immunizations, accompanied by serolog- opportunistic infections, malignancy, and intracranial hem- ical surveillance of women of reproductive age. These pro- orrhage, may lead to brain damage. Thrombocytopenia, grams should be undertaken only if the current expanded from direct damage to the bone marrow or as an indirect programs of immunization are already achieving coverage consequence of opportunistic infections, predisposes HIV- of 80 percent or more. Coverage of less than 80 percent may infected children to intracranial hemorrhages and strokes result in reduced transmission in childhood but leave a large (Mueller 1994). These children are also at higher risk for number of women susceptible when they reach reproduc- opportunistic infections of the central nervous system, such tive age. A recent cost-benefit analysis of universal rubella as toxoplasmosis or cryptococcal meningitis (Aylward et al. vaccination indicates economic benefits comparable to 1992; Mueller 1994), although such infections occur less fre- Haemophilus influenzae type B (Hib) and hepatitis B virus quently in children than in adults. Neurological problems in (HBV) vaccines (Hinman et al. 2002). HIV-infected children have been described with varying frequency from different parts of the world. In a natural his- Congenital Syphilis tory study from South Africa, neurological abnormalities Congenital syphilis is a common and important cause of were found in approximately 50 percent of children fol- diverse clinical manifestations in the newborn infant that lowed to 18 months of age (Bobat et al. 1998). In Rwanda, Developmental Disabilities | 135 15 to 40 percent of HIV-infected infants were found to have In a pooled analysis from five recent studies with similar abnormalities by 6 months of age (Msellati et al. 1993). definitions of cerebral malaria and comparable methodolo- The developmental trajectory of such infected children is gies and diagnostic criteria, a neurological sequelae rate in confounded by maternal, social, and biological risk factors survivors of 16 percent was reported (Snow et al. 1999). during pregnancy and early childhood. Maternal substance A range in reported incidence of neurological sequelae and drug abuse, more common in HIV-infected women, is from 9 percent (Molyneux et al. 1989) to 23 percent known to have an independent adverse effect on brain (van Hensbroek et al. 1997) reflects differences in what are growth and neurodevelopmental outcome. Low birthweight considered to be significant deficits on discharge and those and prematurity, poverty, protein calorie malnutrition, and that resolve over subsequent periods of observation. The micronutrient deficiencies, more frequently seen in HIV- overall rate of persisting neurological sequelae in studies infected children, particularly in developing countries, may that lasted at least six months was 5.6 percent and provides similarly compromise early child development (Brouwers a better estimate of true incidence. Assuming that only chil- et al. 1996). Children who are persistently ill lose a sense of dren with cerebral malaria who reach hospitals are likely to mastery motivation and hence fail to practice new develop- survive, that even these children have a case-fatality rate of ment skills, especially during the first two years of life (Trad 16.7 percent, and that only 5.6 percent of survivors have et al. 1994). Maternal-child interaction is affected as HIV persistent neurological sequelae after six months, it has been disease progresses and as maternal emotional availability estimated that the annual risk of neurological sequelae for decreases, resulting in irregular attachment. The observa- cerebral malaria is 0.03 per 1,000 in children zero to four tion has been made that HIV-uninfected children of HIV- years of age and 0.006 per 1,000 in children five to nine years positive mothers are also at higher risk for cognitive aca- of age (van Hensbroek et al. 1997). This amounts to 2,443 demic and language delays than the general population and 402 annual neurological sequelae events in these two (Condini et al. 1991). This may be similarly mediated age groups, respectively. Because the underlying assump- through the social, economic, and environmental conse- tions for these estimates are quite conservative, it is likely quences of the infection on other household members that the burden of sequelae is higher than the figure pre- (Faithfull 1997; Kotchik 1997; Miles et al. 1997). sented. These assumptions, if correct, also suggest that pri- mary prevention of malaria may have differential impact on malaria-specific mortality and disability. Malaria Malaria is the leading cause of childhood mortality and Bacterial Meningitis morbidity in large tracts of the subcontinent. Because cere- bral malaria is a well-known and not infrequent complica- A recent survey of mostly hospital-based epidemiological tion and may result in neurological sequelae in survivors, studies throughout Sub-Saharan Africa provides an initial malaria has the potential to make a significant contribution basis for calculating incidence rates for bacterial meningitis to the burden of developmental disabilities in Sub-Saharan and its most common pathogens in the subcontinent Africa. (Peltola 2001). These studies also provide important infor- The neuropathology of cerebral malaria stems from a mation on overall and pathogen-specific case-fatality and series of complex mechanisms, which may operate inde- neurological sequelae rates in African children. pendently or in combination to adversely affect the In the absence of prospective population-based studies, developing brain. These include hypoglycemia, multiple the few incidence data available for all causes of bacterial seizures, reduced cerebral perfusion associated with raised meningitis permit a tentative estimate of annual incidence at intracranial pressure, hypoxia associated with microvascu- about 25 per 100,000, or 180,000 cases for the subcontinent lar obstruction, and tissue damage following induction of as a whole. On the basis of published epidemiological and cytokine cascades (Marsh 1995). The sequelae reported laboratory data, 50 percent of cases can be assumed to be in order of frequency were hemiplegia or hemiparesis, caused by Hib, giving an estimate of about 90,000 cases of speech disorders, blindness, hearing impairment, cerebral Hib meningitis per year. Because the vast majority of cases palsy, and epilepsy. Because children often have multiple occurred in children under the age of five years, the incidence neurological sequelae it is not possible to disaggre- in this age group is estimated to be 74 cases per 100,000. gate these data to provide absolute rates for each type of This calculation tallies quite well with data in the same age sequelae. group from five African countries. Annual incidence rates of 136 | Geoff Solarsh and Karen J. Hofman Hib meningitis (per 100,000 children under the age of survivors had significant impairments (Anlar, Yalaz, and five years) was 72 in Senegal (Cadoz, Denis, and Mar 1981), Dizmen 1989; Gadoth et al. 1981). In a recent study in Kenya 62 in Burkina Faso (Tall et al. 1994), 60 in The Gambia significant reductions in head circumference were noted in (Bijlmer and van Alphen 1992), 53 in Niger (Campagne et al. NNT survivors (Barlow et al. 2001). These children also had 1999), and 51 among black children in South Africa (Hussey more problems with hand-eye coordination, lower develop- et al. 1994). Although meningococcal meningitis predomi- mental scores, more mild neurological abnormalities, and nates in the meningitis belt, it has its greatest impact on more frequently reported behavioral problems. adults, and overall, Hib and Streptococcus pneumoniae pre- The main intervention to prevent NNT is the provision dominate as causes of meningitis in young children in the of three doses of tetanus toxoid (TT) to women in their first subcontinent as a whole. pregnancy and a single booster dose in each of the subse- On the basis of data from several studies evaluating neu- quent pregnancies, in addition to following safe hygienic rological sequelae of bacterial meningitis according to etiol- practices at birth and in the postnatal period. In a global ogy, it is estimated that about 40 percent of those surviving end-of-decade assessment of TT coverage, 71.3 percent of Hib meningitis, 50 percent of those surviving pneumococ- pregnant women were reported to have received at least two cal meningitis, and 10 percent of those surviving meningo- doses of TT in their previous pregnancy (UNICEF 1999). It coccal meningitis had long-term sequelae. has been shown in tetanus seroprevalence surveys in Sub- Saharan Africa that estimates of TT coverage significantly under measure protection against NNT (Deming et al. Measles 2002) and that many countries may be approaching World For the vast majority of children with measles infection the Summit of Children year 2000 goals of 90 percent TT cov- main risk is death rather than nonintact survival. A notable erage. This is reflected in a dramatic decline in the number exception, as has been mentioned earlier, is the occurrence of clinical cases in some parts of the subcontinent (Jeena, of blindness in measles survivors, which has been cited as Wesley, and Coovadia 1999). one of the most important preventable causes of blindness Given that vaccination coverage is reasonably high and in the populations of Sub-Saharan Africa. Measles immu- serological protection even higher, and that high mortality nization has long been accepted as one of the most cost- rates in unvaccinated populations result in high neonatal effective interventions in child health. The main challenge mortality, we speculate that NNT makes a very modest con- will be to replicate more widely the high vaccination cover- tribution to burdens of developmental disability in Sub- age and virtual elimination of measles that has been Saharan Africa as a whole. It may represent a more signifi- achieved in many countries on the subcontinent. cant problem in selected countries, where the main thrust must be to improve TT coverage. Tetanus Neonatal tetanus (NNT) remains an important cause of NUTRITION infant mortality in Sub-Saharan Africa, where it has been estimated that 150,000 neonates suffer from NNT each year There is now evidence linking many nutritional deficiencies (Galazka and Gasse 1995). The full extent of the problem to deficits in cognition, motor performance, and behavior. may be much larger than this, and the WHO has suggested There is substantial evidence that protein-energy malnutri- that in some areas as many as 95 percent of cases may go tion and deficiencies in iron and iodine, all of which have unrecognized (WHO 1997). Since high case-fatality rates of been and continue to be prevalent in parts of Sub-Saharan up to 90 percent have been reported for NNT, the main end Africa, are associated with long-term deficits in cognition point of interest in most studies has been neonatal death and school performance. However, this relationship is com- rather than neurological sequelae. In early studies it was plex and affected by the severity and duration of the defi- thought that no permanent neurological damage occurred ciency, the stage of the child's development, the coexistence in survivors of NNT (Gadoth et al. 1981; Sharma et al. of other biological conditions, and a number of sociocultu- 1976). However, frequent uncontrolled spasms associated ral factors. It is difficult to establish that the association is with prolonged apnea and drops in oxygen saturation may causal, as there have been few randomized controlled treat- lead to hypoxic brain damage. Later long-term studies from ment trials with long-term follow-up. The evidence for India and Turkey have suggested that 13 to 37.5 percent of an association between common nutrient deficiencies in Developmental Disabilities | 137 Sub-Saharan Africa and later development is briefly West 1996; WHO 1996). Evidence for the contribution of reviewed below. VAD to the burden of visual impairment in Sub-Saharan Africa is the main interest in this review and has been pre- sented earlier. There is evidence to suggest that there has Iodine Deficiency been a decrease in clinical VAD (ACC/SCN 1997), manifest Iodine deficiency has multiple and serious adverse effects, as eye lesions, largely as a result of more effective vitamin A including impaired cognitive function, and it is considered supplementation programs in many developing countries, to be the leading cause of preventable mental retardation including Sub-Saharan Africa. However, recent estimates for and brain damage worldwide. A 1999 review of data on the clinical VAD and subclinical VAD still range between status of iodine deficiency disorders (IDD) demonstrated 2.8 million and 3.3 million and 140 million to 251 million that IDD is a public health problem in 44 out of 46 coun- preschool children, respectively (UNICEF and Tulane tries in Sub-Saharan Africa and that an estimated 295 mil- University 1998; WHO 1995a). It, therefore, continues to be lion people living in these countries are potentially at risk a major public health problem for which there are relatively for iodine deficiency (WHO, UNICEF, and ICCIDD 1999). simple and cost-effective interventions. Its well-known effect on mental development has played an important role in mobilizing political, public health, and Iron Deficiency nutritional activists in support of national and international prevention programs. Efforts at prevention and control The WHO has estimated that about 40 percent of the focus mainly on the iodization of salt, and as a result of con- world's population suffer from anemia and that a substan- certed public health prevention efforts, salt iodization was tial proportion of that burden is attributable to iron defi- reported, by 1999, to have reached 63 percent of households ciency. There is a particularly high prevalence of anemia in in Sub-Saharan Africa. In a recent study of seven African pregnant women (50 percent), children in the first two years countries one year following the introduction of salt iodiza- of life (40 percent), and schoolchildren (40 percent) tion, increases in median urinary iodine to above levels con- (WHO 2000). With regard to preschool children, anemia sidered to constitute iodine deficiency was observed in all prevalence is the highest in Africa (42 to 53 percent) and countries (Delange, de Benoist, and Alnwick 1999). Progress Asia of all WHO regions (WHO 1998). An association toward the elimination of IDD through universal salt between hemoglobin concentration and psychomotor per- iodization, with its anticipated impact on cognitive impair- formance has been demonstrated at all stages of life. ment, appears to be one of the most significant successes in Although there is a good biological basis for claiming that a the field of noncommunicable diseases. However, the chal- deficiency of iron might impair mental and motor develop- lenge still remains to ensure that salt iodization reaches all ment, an Expert Group, in weighing up the available evi- populations. Recent studies in South Africa and Lesotho dence (Idjradinata and Pollitt 1993), concluded that only showed that, in spite of compulsory iodization of salt, anemia, and not iron deficiency without anemia, impairs iodine deficiency remains a significant problem in primary the behavior and development of infants (Draper 1997). school children in more remote areas of the country There now appears to be sufficient evidence to show that (Sebotsa et al. 2003; van Stuijvenberg et al. 1999). iron supplementation of anemic children over two years of age improves development but that this effect is less conclu- sive in children under the age of two years (Grantham- Vitamin A Deficiency McGregor and Ani 2001). Vitamin A is an essential micronutrient for normal growth, for normal functioning of the visual system, for the mainte- Poverty and Protein-Energy Malnutrition nance of normal epithelial integrity and immune function, and for normal reproduction. Consequently, vitamin A Although there is evidence that the levels of protein-energy deficiency (VAD) results in increased severity of certain malnutrition (PEM) in Sub-Saharan populations have infections and an increased risk of disease and death in improved and the spectrum of these deficiencies has young children. More severe vitamin A depletion leads to changed over the past 30 to 40 years, there is still a substan- night blindness, which can evolve to irreversible partial or tial burden of PEM in many populations in the subregion. total blindness if the depletion continues (Sommer and For many reasons it has been difficult to establish a causal 138 | Geoff Solarsh and Karen J. Hofman relationship between undernutrition and behavioral and (Grantham-McGregor 1995). In unraveling the complex cognitive development in these children. The main classifi- relationship between malnutrition and cognition, several cation of malnutrition used in these studies defines a questions remain unanswered--these include questions mixture of clinical signs that are a product of coexistent about the duration of the cognitive deficits, the specific infections and many other deficiencies, such as zinc, magne- types of cognitive functions that are affected, the relative sium, copper, and iron, as well as protein and energy, each of effect of malnutrition at different points in early childhood, which may have an independent and different effect on whether complete recovery can occur following placement developmental outcomes (Grantham-McGregor 1995). in enriched environments, and the relative contribution of Malnourished children usually come from families who individual nutrients to cognitive deficits. suffer from numerous disadvantages. These include poor social, economic, and environmental living conditions and unstable family units with large numbers of closely spaced ENVIRONMENTAL TOXINS children. The parents are often unwell, poorly nourished, Alcohol and depressed; young with low intelligence and levels of education; either unemployed or in low-skilled occupations; Fetal alcohol syndrome is the most common single prevent- and likely to have low social and media contacts. Few toys or able cause of mental retardation worldwide (Viljoen 1999), books can be found in the homes of these families, and par- and the overall rate for the developed world, where the vast ents participate little in play activities; thus there is little majority of prevalence studies have been based, has been stimulation. placed at 0.97 per 1,000 children (Abel 1995). It has its high- The best evidence for a causal effect of undernutrition on est rates in subgroups of the population characterized by cognition and behavior is likely to come from randomized low socioeconomic status and confounded by race. For controlled trials. It has been considered unethical to con- example, in American Indians and African Americans, rates duct randomized controlled trials to study the effects of of 8 per 1,000 children (May et al. 1983) and 2.29 per 1,000 malnutrition on behavior and development. It has therefore have been found, considerably in excess of rates in the U.S. been necessary to rely on less satisfactory epidemiological population as a whole. study designs, such as case-control studies, which have been The only published studies or reports on fetal alcohol unable to control for the wide-ranging nature of these syndrome in Sub-Saharan Africa are confined to a subgroup children's disadvantages. of the "coloured," or mixed race, community in the Western There is also some uncertainty about whether cognitive Cape of South Africa, where a prevalence rate of 40.5 to impairment is an inevitable consequence of severe early 46.4 per 1,000 children age five to nine years has recently malnutrition and what form it is likely to take. Animal been measured (May et al. 2000). This is the highest rate for experimentation initially suggested that early malnutrition fetal alcohol syndrome ever recorded anywhere in the world significantly reduced brain growth and left it permanently and appears to be attributable to very high alcohol intake in smaller in size (Winick and Noble 1966). More recent work this South African subpopulation as a consequence of a spe- has shown that many of these dramatic anatomic alterations cial set of historical and social conditions. are reversible. It also shows that parallel alterations in neu- The principal effect of alcohol teratogenicity is brain rotransmitter and receptor characteristics are evident, damage with consequent lowering of intelligence (mean resulting in subtler neurodevelopmental deficits in motiva- intelligence quotient 65), behavioral abnormalities (atten- tion, emotional reactivity, and cognitive flexibility rather tion deficit, hyperactivity, aggressiveness), and poor lan- than in frank reductions in intelligence (Levitsky and guage assimilation. The syndrome also presents with a rec- Strupp 1995). ognizable cluster of facial features and a maternal history of In spite of all these limitations and uncertainties evidence heavy alcohol intake in pregnancy, often imbibed in a pat- suggests that previously malnourished children show a tern of binge drinking. There is no reason to believe that the deficit in tests of cognitive function or intelligence. This overall prevalence in Sub-Saharan Africa is particularly effect is particularly strong if the undernutrition is chronic high, but it may occur in other subpopulations with low and if they return to poor environments. The effect is less socioeconomic status and histories of high alcohol intake. In clear in children exposed to acute episodes of malnutrition these subpopulations, interventions to prevent or reduce who do not return to impoverished social conditions alcohol ingestion should assume high priority. Developmental Disabilities | 139 Other Toxins (Mock et al. 1999). The majority of these injuries in the zero to four age group were due to falls or burns in both urban Other toxins, such as drugs, nicotine, and heavy metals, may and rural populations. result in cognitive and developmental deficits in young chil- dren. No information could be found on the prevalence of disabilities attributable to these toxins in Sub-Saharan INTERVENTIONS Africa, although there is clear evidence of exposure to lead, for example, and every reason to expect that some contribu- Given the large and continued contribution of infectious, tion, albeit small, is made by these and other toxins to the perinatal, and nutritional conditions to under-five mortality burden of disabilities in this subregion. in many parts of Sub-Saharan Africa, the goal and measure of success of most public child-health interventions in Sub-Saharan Africa continues to be an improvement in ACCIDENTS AND INJURIES child survival. We argued earlier that since the majority of these children (90 to 95 percent) are likely to survive beyond For children in Africa who survive the first four years of life, early childhood, some consideration should be given to the head injury becomes the most likely cause of disability or reduction of developmental disabilities as a parallel end death, and this remains true until the fourth decade of life point for child health programs in the subcontinent. We also (Kibel, Joubert, and Bradshaw 1990). Insofar as these acci- argued that given the increasing focus on pregnant women dents and injuries result in significant damage to the devel- and neonates as the target groups for these interventions oping central nervous systems, they can be said to make a and shared causal pathways for mortality and disability in contribution to the burden of developmental disabilities in early childhood, these parallel end points, namely, reduc- young children in the subcontinent. In 1990, DALY rates tions in mortality and developmental disabilities, make pro- attributable to injuries and noncommunicable diseases, grammatic sense. taken together, among children age five to fourteen years The effect of existing interventions that reduce child had already exceeded those attributable to infectious, peri- mortality on the prevalence and incidence of childhood dis- natal, and nutritional conditions (Deen et al. 1999). DALY abilities is likely to be complex. In the absence of any data on rates attributable to injuries were highest in Sub-Saharan this relationship in developing countries, it seems reason- Africa and India, were higher for boys than for girls and for able to speculate that effective primary prevention of many children zero to four years than for children five to nine of the conditions covered in this review, such as vaccine- years of age. Road traffic accidents, falls, burns, and acci- preventable diseases, is likely to have parallel effects in dental poisoning are the most common categories of child- reducing burdens of both mortality and disability in Sub- hood accidents and unintentional injuries. The largest num- Saharan Africa, although the extent of disability reduction ber of intentional injuries was caused by war. Hundreds of may be quite modest. It has been suggested that this group thousands of children are permanently disabled in Africa of children in developing countries may have a dispropor- every year as a consequence of war injuries and poor trauma tionately high mortality rate later in childhood as a conse- care (Bickler and Rode 2002). During 2000 a total of 11 quence of ongoing limitations in their access to high-quality major wars were being fought in Africa, involving 20 percent medical and social care (Durkin 2002). of the population on the subcontinent. It is estimated that Secondary prevention is likely to have, if anything, the 120,000 to 200,000 child soldiers age five to sixteen years are opposite effect; intervention at the point where disease is participating in such conflicts, some of whom sustain bullet already present may prevent death at the expense of nonin- and shrapnel wounds as well as burns and land mine tact survival. Interestingly, although cerebral palsy preva- injuries. No figures are available for intentional injuries, but lence increased during the early 1970s and 1980s in devel- these figures may significantly increase the total burden of oped countries (Bhushan, Paneth, and Kiely 1993; Stanley injuries in some populations. and Watson 1992), cerebral palsy rates in the 1980s and In a community-based survey of injury-related disability 1990s, presumably as a consequence of improvements in in Ghana, children zero to five and five to fourteen years of neonatal intensive care and increased survival of infants age accounted for 14 percent and 10 percent, respectively, with antenatal and postnatal brain damage, have been either of all disability in urban populations and 2.8 percent and stable (Colver et al. 2000; Hagberg et al. 1996) or decreas- 12.3 percent, respectively, of injuries in rural populations ing (Grether and Nelson 1997; O'Shea, Klinepeter, and 140 | Geoff Solarsh and Karen J. Hofman Dillard 1998; Topp, Uldall, and Langhoff-Roos 1997) as sur- developmental disabilities. Together these data provide vival rates continue to increase. It has been suggested that the necessary information to formulate potentially viable changing rates of postnatal brain damage are the more like- interventions and to set up efficacy trials to test them. ly explanation for the secular trends in cerebral palsy preva- Subsequent steps, of less immediate relevance to this chap- lence in developed countries over the last three decades ter, involve the refinement of these interventions based (O'Shea 2002). These trends are confounded by variations on trial findings and their application and evaluation in in case definitions, ascertainment strategies, and study "real-world" settings. methods between studies and over time. These issues, par- ticularly in the area of secondary prevention, will be equally Secondary Prevention if not more problematic for the measurement of trends in developmental disability prevalence in developing countries Neonatal screening programs for congenital disorders, the and particularly in Sub-Saharan Africa. identification of high-risk newborns for early multidiscipli- nary intervention programs, or screening in infancy for developmental delays at primary health care clinics have all Primary Prevention been proposed as secondary prevention strategies to mini- Population-based interventions aimed at primary preven- mize the effect of neurological disorders once they have tion of diseases the risk of which is attributable to high already occurred. In some Sub-Saharan countries, such as population offer the most attractive approach to tackling South Africa, that have high attendance for antenatal care, developmental disabilities in Sub-Saharan Africa. A general supervised deliveries, postnatal care, and "well-baby" servic- conceptual framework that has been proposed to guide the es in the first year of life, these strategies provide the possi- development and evaluation of all public health interven- bility of early intervention within established health facility tions (de Zoysa et al. 1998) has potential applications for networks at strategic points during the neonatal and post- interventions in the field of developmental disabilities neonatal periods. Once screening has occurred, they permit (figure 10.4). definitive assessments and the institution of appropriate The starting point in this framework is a detailed descrip- programs of management for identified children. The feasi- tion of the prevalence, types, severity, and distribution of bility and relative benefit of these access strategies and early developmental disabilities in methodologically sound two- intervention programs need to be carefully weighed against phase population-based surveys. The case for good des- primary and tertiary prevention in different population criptive epidemiological data on developmental disabilities settings. in Sub-Saharan Africa has already been made. Given the A full consideration of the available primary and second- expense and effort in mounting these surveys, these descrip- ary interventions and their relative cost-effectiveness is tions have often been combined in previous studies with the beyond the scope of this chapter. An exercise of this kind is next steps in the research framework, namely, the determi- under way and will be the subject of a chapter in a separate nation of biological risk factors and social determinants for monograph (Disease Control Priorities in Developing Figure 10.4 Causal Pathways for Developmental Disabilities Step 1 Step 2 Step 3 Step 4 Step 5 Identify Explore Test Formulate Describe the biological and context and interventions possible problem behavioral identify in rigorous interventions risk factors determinants efficacy trials Step 9 Step 8 Step 7 Step 6 Monitor Assess the Assess the Formulate impact of effectiveness efficacy of public health intervention of public public health interventions at scale health interventions interventions Source: De Zoysa et al. 1998. Developmental Disabilities | 141 Countries, 2nd ed.) Because tertiary prevention is concerned of opportunities and social integration of people with with ameliorating the impact of established developmental disabilities (WHO, UNESCO, and ILO 1994). CBR activities disabilities and has been an important, if not the main, are holistic in nature and, in attempting to integrate social thrust in Sub-Saharan Africa, it merits special attention. and medical models of disability, include all or some of the following: awareness raising in communities; advocacy for the rights of people with disabilities; development of par- Tertiary Prevention ents' and caregivers' groups; income generation; networking Tertiary prevention programs are needed for all children with educational, social, and employment authorities; and with established disabilities and this is particularly true for provision of rehabilitation and health services. The focal Sub-Saharan Africa, where a large burden of established person in the CBR program is the person with disabilities, disability already exists as a result of past failures in primary and the main goals, in addition to providing access to reha- and secondary prevention of the many conditions discussed bilitation services, are to ensure that people with disabilities earlier in this chapter. As the focus turns from mortality to have rights to self-determination; enjoy the full benefits of disability as a public health end point, a better understand- family membership; are active and responsible community ing is needed of the dynamic interplay between existing child members; have equal access to education, skills training, survival strategies and interventions to specifically reduce or work, and recreation; play a significant role in the CBR pro- ameliorate developmental disabilities in developing coun- gram itself; participate in organizations that cater to people tries, in general, and in Sub-Saharan Africa, in particular. with disabilities; and act as lobbyists for the disabled and For the majority of countries in Sub-Saharan Africa, their families. service provision to people with established disabilities will CBR programs vary in their content, but most include be the starting point and, for some time, the dominant some of the following services: approach in addressing the needs of people with develop- mental disabilities. The concern in these programs is less · a decentralized approach to service delivery with the with underlying causes than with improvement of primary focus and integration of these services at dis- functional capacity of affected individuals and the enhance- trict level within established models of primary health ment of their participation in all aspects of community life. care Their aim will be to combine direct and indirect therapeu- · training and support for general-purpose mid-level reha- tic inputs by health and rehabilitation workers with support bilitation workers who operate at household and com- and training for families and caregivers and societal inter- munity level ventions to limit stigma and equalize opportunities in all · screening and early identification of children with devel- walks of life. In all these processes people with disabilities opmental disabilities and their immediate family members will be expected to · home-based support and training for activities of daily take a leading role with support and guidance, but not direc- living tion, from health and rehabilitation professionals. · promotion of inclusive education in pre- and primary For health and rehabilitation professionals in developing schools countries faced with the large number of children and · provision of vocational training at secondary schools adults with established disability, tertiary prevention contin- · referral to specialist rehabilitation services ues to remain an important priority. The failure to meet the · reorientation of the roles of rehabilitation professionals, rehabilitation needs of the majority of people with disabili- where available, to support, train, and manage rehabilita- ties through existing services and growing concerns about tion teams and CBR programs in addition to directly the inherent limitations of institutional rehabilitation gave providing rehabilitation services. rise to an alternative approach to care for the disabled in developing countries, called community-based rehabilita- tion (CBR). Defining the sequence and level of input at each of these levels may vary from country to country and from program to program but will be facilitated in all cases by the existence Community-Based Rehabilitation of national health systems based on primary health care and The concept of CBR has been defined by the WHO as a national rehabilitation strategies at all tiers of the health community development strategy focusing on equalization system. 142 | Geoff Solarsh and Karen J. Hofman EVALUATION AND COST-BENEFIT ANALYSES CONCLUSION Research and evaluation has not been a prominent feature There is an extreme dearth of good-quality data on the of the CBR movement. These programs occur in resource- prevalence, types, and causes of developmental disabilities constrained environments, where the need for service pro- in Sub-Saharan Africa. Although a fair number of facility- vision is preeminent and the necessary motivation, based and population-based studies do provide some epi- resources, and skills for research have often been lacking. demiological descriptions of the conditions that together However, many questions remain about CBR, and the give rise to the childhood burden of disability, these data field is criticized for having poor indicators with which to vary in quality and there is insufficient understanding of measure success (Wirz and Thomas 2002). Divergent their relative and population-attributable risks for disability. views about the primary goals and end points of CBR This makes it difficult to quantify disability burdens for chil- programs have to some extent impeded the development dren on the subcontinent with accuracy and to intervene of a clear set of indicators. Although some believe that effectively in these conditions or to advocate on behalf of standard end points, such as measures of functional children with developmental disabilities. Studies are improvement of specific disabilities, are required, others urgently needed that provide these data and, as the health are more concerned with measuring shifts in attitudes of transition deepens in certain countries, that focus on dis- community members toward people with disabilities or ability rather than mortality as an end point. the extent of inclusion of the disabled in institutional or Developmental disabilities bring together clinicians from community life. Those who see CBR as a community medical, nursing, and rehabilitation backgrounds; scientists development activity are interested in the direct participa- with biomedical, population, and social science perspec- tion of people with disabilities in CBR program activities tives; and researchers, service providers, and people with or income-generating activities as key measures of disabilities with quite different worldviews. Together they outcome. Most practitioners, however, see all of these out- must tackle this public health problem through balanced comes as important. An approach to evaluation that approaches to primary, secondary, and tertiary prevention. integrates medical and social models is needed, and the The immediate challenge in countries with large compet- recently revised WHO International Classification of ing health needs and severe resource constraints may be to Functioning, discussed earlier in this chapter, may provide "start at the end." Many underserved populations in these useful and more standardized guidelines for future countries carry large and unaddressed burdens of estab- approaches of this kind. lished disabilities. The provision of services and programs A review of the literature confirms that there is a dearth that aim to improve the function of affected children and of rigorous evaluations of CBR programs. This is probably improve their participation in these societies is their right as much a reflection of the lack of well-designed programs and our obligation, and this goal must be pursued in paral- as it is of the lack of well-designed evaluations, because the lel with strategies to tackle primary and secondary preven- two processes often go hand-in-hand. There appears to be tion of the many conditions that contribute to the burden an urgent need for the delineation of an evaluation frame- of disabilities in the subcontinent. The diverse and func- work that lays out the goals of CBR programs and clearly tional needs of these children in these constrained environ- characterizes the possible inputs, outputs, and expected ments must be analyzed, and appropriate programs that outcomes of these programs in order both to inform future realistically address these needs and constraints must be program design and to guide subsequent program developed and carefully evaluated. For many children these evaluation. needs, unlike their basic health needs, are simply not being Given that tertiary prevention will continue to be an met. Without demonstrating care for the affected, health important intervention for developmental disabilities, irre- workers may lack the credibility to advocate for prevention spective of the successes of primary and secondary preven- of disabilities in the unaffected. tion, and that CBR programs are likely to continue to be the strategy of choice in developing countries, there is an urgent need for credible evaluations that show an advantageous REFERENCES cost-benefit relationship. Without this it will be increasingly Abel, E. L. 1995. "An Update on Incidence of FAS: FAS Is Not an Equal difficult to obtain the ongoing support of policy makers and Opportunity Birth Defect." Neurotoxicology and Teratology 17 (4): external funders for these programs. 437­43. Developmental Disabilities | 143 ACC/SCN (Administrative Committee on Coordination/Sub-committee Christianson, A. L., and J. G. Kromberg. 1996. "Maternal Non-Recognition on Nutrition). 1997. Third Report on the World Nutrition Situation. of Down Syndrome in Black South African Infants." Clinical Genetics Geneva: ACC/SCN. 49 (3): 141­44. Adeoye, A. 1973. "Sickle-Cell Anaemia." British Medical Journal 2 (5861): Christianson, A. L., M. E. Zwane, P. Manga, E. Rosen, A. Venter, D. Downs, 304. and J. G. Kromberg. 2002. "Children with Intellectual Disability in ------. 1996. "Survey of Blindness in Rural Communities of South- Rural South Africa: Prevalence and Associated Disability." Journal of Western Nigeria." Tropical Medicine and International Health 1 (5): Intellectual Disability Research 46 (Part 2): 179­86. 672­76. Cleland, J., N. Onuoha, and I. Timaeus. 1994. "Fertility Change in Sub- Adeyokunnu, A. A. 1982. "The Incidence of Down's Syndrome in Nigeria." Saharan Africa: A Review of the Evidence." In The Onset of the Fertility Journal of Medical Genetics 19 (4): 277­79. Transition in Sub-Saharan Africa, ed. T. Locoh and V. Hertrich. Liège: International Union for the Scientific Study of Population. Ahmad, O. B., A. D. Lopez, and M. Inoue. 2000. "The Decline in Child Mortality: A Reappraisal." Bulletin of the World Health Organization 78 Cohen, B. 1993. "Fertility Levels, Differentials and Trends." In (10): 1175­91. Demographic Change in Sub-Saharan Africa, ed. K. Foote, K. H. Hill, and L. G. Martin, 8­67. Washington DC: National Academy Press. Alwan, A., and B. Modell. 1997. Community Control of Genetic and Colver, A. F., M. Gibson, E. N. Hey, S. N. Jarvis, P. C. Mackie, and S. Congenital Disorders. EMRO Technical Publications Series 24. Geneva: Richmond. 2000. "Increasing Rates of Cerebral Palsy Across the WHO Regional Office for the Eastern Mediterranean. Severity Spectrum, in North-east England 1964­1993. The North of Anand, S., and K. Hanson. 1997. "Disability-Adjusted Life Years: A Critical England Collaborative Cerebral Palsy Survey." Archives of Disease in Review." Journal of Health Economics 16 (6): 685­702. Childhood. Fetal and Neonatal Edition 83 (1): F7­F12. Anlar, B., K. Yalaz, and R. Dizmen. 1989. "Long-Term Prognosis after Condini, A., G. Axia, C. Cattelan, M. R. D'Urso, A. M. Laverda, F. Viero, Neonatal Tetanus." Developmental Medicine and Child Neurology 31 and F. Zacchello. 1991. "Development of Language in 18­30-Month- (1): 76­80. Old HIV-1-Infected but Not Ill Children." AIDS 5 (6): 735­39. Axton, J. H., and L. F. Levy. 1974. "Mental Handicap in Rhodesian African Cornielje, H., P. Ferrinho, D. Coetsee, and S. G. Reinach. 1993. Children." Developmental Medicine and Child Neurology 16 (3): 350­55. "Development of a Community-based Rehabilitation Programme for Aylward, E. H., A. M. Butz, N. Hutton, M. L. Joyner, and J. W. Vogelhut. a Poor Urban Area in South Africa. A Disability Prevalence Study." 1992. "Cognitive and Motor Development in Infants at Risk for Community Health Association of South Africa Journal 14 (1): 26­32. Human Immunodeficiency Virus." American Journal of Diseases in Couper, J. 2002. "Prevalence of Childhood Disability in Rural Kwazulu- Children 146 (2): 218­22. Natal." South African Medical Journal 92 (7): 549­52. Barlow, J. L., V. Mung'Ala-Odera, J. Gona, and C. R. Newton. 2001. "Brain Cutts, F. T., S. E. Robertson, J. L. Diaz-Ortega, and R. Samuel. 1997. Damage after Neonatal Tetanus in a Rural Kenyan Hospital." Tropical "Control of Rubella and Congenital Rubella Syndrome (CRS) in Medicine and International Health 6 (4): 305­8. Developing Countries. Part 1: Burden of Disease from CRS." Bulletin of Bhushan, V., N. Paneth, and J. L. Kiely. 1993. "Impact of Improved Survival the World Health Organization 75 (1): 55­68. of Very Low Birth Weight Infants on Recent Secular Trends in the Cutts, F. T., and E. Vynnycky. 1999."Modelling the Incidence of Congenital Prevalence of Cerebral Palsy." Pediatrics 91 (6): 1094­1100. Rubella Syndrome in Developing Countries." International Journal of Bickler, S. W., and H. Rode. 2002. "Surgical Services for Children in Epidemiology 28 (6): 1176­84. Developing Countries." Bulletin of the World Health Organization 80 de Zoysa, I., J. P. Habicht, G. Pelto, and J. Martines. 1998. "Research Steps (10): 829­35. in the Development and Evaluation of Public Health Interventions." Bijlmer, H. A., and L. van Alphen. 1992. "A Prospective, Population-Based Bulletin of the World Health Organization 76 (2): 127­33. Study of Haemophilus Influenzae Type B Meningitis in The Gambia Deen, J. L., T. Vos, S. R. Huttly, and J. Tulloch. 1999. "Injuries and and the Possible Consequences." Journal of Infectious Diseases 165 Noncommunicable Diseases: Emerging Health Problems of Children (Suppl. 1): S29­32. in Developing Countries." Bulletin of the World Health Organization 77 Bobat, R., D. Moodley, A. Coutsoudis, H. Coovadia, and E. Gouws. 1998. (6): 518­24. "The Early Natural History of Vertically Transmitted HIV-1 Infection Delange, F., B. de Benoist, and D. Alnwick. 1999. "Risks of Iodine-Induced in African Children from Durban, South Africa." Annals of Tropical Hyperthyroidism after Correction of Iodine Deficiency by Iodized Paediatrics 18 (3): 187­96. Salt." Thyroid 9 (6): 545­56. Breart, G., and A. S. Poisson-Salomon. 1988. "Intrauterine Growth Deming, M. S., J. B. Roungou, M. Kristiansen, I. Heron, A. Yango, A. Retardation and Mental Handicap: Epidemiological Evidence." Guenengafo, and R. Ndamobissi. 2002."Tetanus Toxoid Coverage as an Baillieres Clinical Obstetrics and Gynaecology 2 (1): 91­100. Indicator of Serological Protection against Neonatal Tetanus." Bulletin Brouwers, P., C. Decarli, M. P. Heyes, H. A. Moss, P. L. Wolters, G. Tudor- of the World Health Organization 80 (9): 696­703. Williams, L. A. Civitello, et al. 1996. "Neurobehavioral Manifestations Draper, A. 1997. Child Development and Iron Deficiency: The Oxford Brief. of Symptomatic HIV-1 Disease in Children: Can Nutritional Factors Opportunities for Micronutrient Interventions Project and Partnership Play a Role?" Journal of Nutrition 126 (Suppl. 10): 2651S­62S. for Child Development. Washington, DC: USAID. Brustle, O., H. Spiegel, S. L. Lieb, T. Finn, H. Stein, P. Kleihues, and Drugan, A., Y. Yaron, R. Zamir, S. A. Ebrahim, M. P. Johnson, and M. I. O. D. Wiestler. 1992. "Distribution of Human Immunodeficiency Evans. 1999. "Differential Effect of Advanced Maternal Age on Prenatal Virus (HIV) in the CNS of Children with Severe HIV Diagnosis of Trisomies 13, 18, and 21." Fetal Diagnosis and Therapy 14 Encephalomyelopathy." Acta Neuropathology (Berl) 84 (1): 24­31. (3): 181­84. Cadoz, M., F. Denis, and I. D. Mar. 1981. "An Epidemiological Study of Durkin, M. 2002. "The Epidemiology of Developmental Disabilities in Purulent Meningitis Cases Admitted to Hospital in Dakar, 1970­1979." Low-income Countries." Mental Retardation and Developmental Bulletin of the World Health Organization 59 (4): 575­84. Disabilities Research Reviews 8 (3): 206­11. Campagne, G., A. Schuchat, S. Djibo, A. Ousseini, L. Cisse, and J. P. Durkin, M. S., L. L. Davidson, P. Desai, Z. M. Hasan, N. Khan, P. E. Shrout, Chippaux. 1999. "Epidemiology of Bacterial Meningitis in Niamey, M. J. Thorburn, et al. 1994. "Validity of the Ten Questions Screened for Niger, 1981­96." Bulletin of the World Health Organization 77 (6): Childhood Disability: Results from Population-Based Studies in 499­508. Bangladesh, Jamaica, and Pakistan." Epidemiology 5 (3): 283­89. 144 | Geoff Solarsh and Karen J. Hofman Durkin, M. S., Z. M. Hasan, and K. Z. Hasan. 1995. "The Ten Questions Hussey, G., J. Hitchcock, H. Schaaf, G. Coetzee, D. Hanslo, E. van Screen for Childhood Disabilities: Its Uses and Limitations in Schalkwyk, J. Pitout, et al. 1994. "Epidemiology of Invasive Pakistan." Journal of Epidemiology and Community Health 49 (4): Haemophilus Influenzae Infections in Cape Town, South Africa." 431­36. Annals of Tropical Paediatrics 14 (2): 97­103. ------. 1998. "Prevalence and Correlates of Mental Retardation among Idjradinata, P., and E. Pollitt. 1993. "Reversal of Developmental Delays in Children in Karachi, Pakistan." American Journal of Epidemiology Iron-Deficient Anaemic Infants Treated with Iron." Lancet 341 (8836): 147 (3): 281­88. 1­4. Faithfull, J. 1997. "HIV-Positive and AIDS-Infected Women: Challenges Irlam, J. 1996. "The Prevalence of Childhood Disability in a Rural Area of and Difficulties of Mothering." American Journal of Orthopsychiatry 67 KwaZulu-Natal." Amatikulu Primary Health Care Training Centre, (1): 144­51. Durban. Fauveau, V., B. Wojtyniak, J. Chakraboty, A. M. Sarder, and A. Briend. 1990. Islam, S., M. S. Durkin, and S. S. Zaman. 1993. "Socioeconomic Status "The Effect of Maternal and Child Health and Family Planning and the Prevalence of Mental Retardation in Bangladesh." Mental Services on Mortality: Is Prevention Enough?" British Medical Journal Retardation 31 (6): 412­17. 301 (6743): 103­7. Jeena, P. M., A. G. Wesley, and H. M. Coovadia. 1999. "Admission Fryers, T. 1992. "Epidemiology and Taxonomy in Mental Retardation." Patterns and Outcomes in a Paediatric Intensive Care Unit in South Paediatric and Perinatal Epidemiology 6 (2): 181­92. Africa Over a 25-Year Period (1971­1995)." Intensive Care Medicine 25 (1): 88­94. Gadoth, N., R. Dagan, U. Sandbank, D. Levy, and S. W. Moses. 1981. "Permanent Tetraplegia as a Consequence of Tetanus Neonatorum. Jelsma, J., W. De Weerdt, and P. De Kock. 2002. "Disability Adjusted Life Evidence for Widespread Lower Motor Neuron Damage." Journal of Years (DALYs) and Rehabilitation." Disability and Rehabilitation 24 (7): Neurological Sciences 51 (2): 273­78. 378­82. Galazka, A., and F. Gasse. 1995."The Present Status of Tetanus and Tetanus Katzenellenbogen, J., G. Joubert, K. Rendall, and T. Coetzee. 1995. Vaccination." Current Topics in Microbiology and Immunology 195: "Methodological Issues in a Disablement Prevalence Study: Mitchells 31­53. Plain, South Africa." Disability and Rehabilitation 17 (7): 350­57. Gilbert, C. E., L. Anderton, L. Dandona, and A. Foster. 1999. "Prevalence of Kibel, S. M., G. Joubert, and D. Bradshaw. 1990. "Injury-Related Mortality Visual Impairment in Children: A Review of Available Data." in South African Children, 1981­1985." South African Medical Journal Ophthalmic Epidemiology 6 (1): 73­82. 78 (7): 398­403. Kotchik, B. 1997. "The Impact of Maternal HIV Infection on Parenting in Gilbert, C. E., R. Canovas, M. Hagan, S. Rao, and A. Foster. 1993. "Causes Inner-City African American Families." Journal of Family Psychology of Childhood Blindness: Results from West Africa, South India and 11: 447­61. Chile." Eye 7 (Pt. 1): 184­88. Levitsky, D. A., and B. J. Strupp. 1995. "Malnutrition and the Brain: Gilbert, C. E., M. Wood, K. Waddel, and A. Foster. 1995. "Causes of Changing Concepts, Changing Concerns." Journal of Nutrition 125 Childhood Blindness in East Africa: Results in 491 Pupils Attending (Suppl. 8): 2212S­20S. 17 Schools for the Blind in Malawi, Kenya, and Uganda." Ophthalmic Epidemiology 2 (2): 77­84. Luzzato, L., and A. Mehta 1989. "Glucose-6-Phosphate Dehydrogenase Deficiency." In The Metabolic Basis of Inherited Disease, 6th ed., ed. Gloyd, S., S. Chai, and M. A. Mercer. 2001. "Antenatal Syphilis in Sub- C. R. Scriver, A. L. Beaudet, W. S. Sly, and D. Valle. New York: Saharan Africa: Missed Opportunities for Mortality Reduction." Health McGraw-Hill. Policy and Planning 16 (1): 29­34. MacFarlane, A., and M. Mugford. 1984. "Birth Counts." In Statistics of Goldenberg, R. L., G. R. Cutter, H. J. Hoffman, J. M. Foster, K. G. Nelson, Pregnancy and Childhood. London: Her Majesty's Stationery Office. and J. C. Hauth. 1989. "Intrauterine Growth Retardation: Standards for Diagnosis." American Journal of Obstetrics and Gynecology 161 (2): Manni, J. J., and P. N. Lema. 1987. "Otitis Media in Dar es Salaam, 271­77. Tanzania." Journal of Laryngology and Otology 101 (3): 222­28. Grant, J. 1992. State of the World's Children. New York: UNICEF. Marsh, K. 1995. "Neurological Sequelae of Cerebral Malaria." Paper pre- sented at the WHO/TDR meeting, Geneva, December 4­8. Grantham-McGregor, S. 1995. "A Review of Studies of the Effect of Severe May, P. A., L. Brooke, J. P. Gossage, J. Croxford, C. Adnams, K. L. Jones, Malnutrition on Mental Development." Journal of Nutrition 125 L. Robinson, et al. 2000. "Epidemiology of Fetal Alcohol Syndrome in (Suppl. 8): 2233S­38S. a South African Community in the Western Cape Province." American Grantham-McGregor, S., and C. Ani 2001. "A Review of Studies on the Journal of Public Health 90 (12): 1905­12. Effect of Iron Deficiency on Cognitive Development in Children." May, P. A., K. J. Hymbaugh, J. M. Aase, and J. M. Samet. 1983. Journal of Nutrition 131 (2S­2): 649S­66S, discussion 666S­68S. "Epidemiology of Fetal Alcohol Syndrome among American Indians of Grether, J. K., and K. B. Nelson. 1997. "Maternal Infection and Cerebral the Southwest." Social Biology 30 (4): 374­87. Palsy in Infants of Normal Birth Weight." Journal of the American McPherson, B., and C. A. Holborow. 1985. "A Study of Deafness in West Medical Association 278 (3): 207­11. Africa: The Gambian Hearing Health Project." International Journal of Hagberg, B., G. Hagberg, I. Olow, and L. van Wendt. 1996. "The Changing Pediatric Otorhinolaryngology 10 (2): 115­35. Panorama of Cerebral Palsy in Sweden. VII. Prevalence and Origin in Milaat, W. A., T. M. Ghabrah, H. M. Al-Bar, B. A. Abalkhail, and M. N. the Birth Year Period 1987­1990." Acta Paediatrica 85 (8): 954­60. Kordy. 2001. "Population-Based Survey of Childhood Disability in Hatcher, J., A. Smith, I. Mackenzie, S. Thompson, I. Bal, I. Macharia, P. Eastern Jeddah Using the Ten Questions Tool." Disability and Mugwe, et al. 1995. "A Prevalence Study of Ear Problems in School Rehabilitation 23 (5): 199­203. Children in Kiambu District, Kenya, May 1992." International Journal Miles, M., P. Burchinal, D. Holditch-Davis, Y. Wasilewski, and B. Christian. of Pediatric Otorhinolaryngology 33 (3): 197­205. 1997. "Personal, Family and Health-Related Correlates of Depressive Hill, K., R. Pande, M. Mahy, and G. Jones. 1999. Trends in Child Mortality Symptoms in Mothers with HIV." Journal of Family Psychology 11 (1): in the Developing World: 1966­1996. New York: UNICEF. 23­34. Hinman, A. R., B. Irons, M. Lewis, and K. Kandola. 2002. "Economic Mock, C. N., F. Abantanga, P. Cummings, and P. Koepsell. 1999. "Incidence Analyses of Rubella and Rubella Vaccines: A Global Review." Bulletin of and Outcome of Injury in Ghana: A Community-Based Survey." the World Health Organization 80 (4): 264­70. Bulletin of the World Health Organization 77 (12): 955­64. Developmental Disabilities | 145 Molteno, C., and P. Lachman. 1996. "The Aetiology of Learning Disability Sebotsa, M. L., A. Dannhauser, P. L. Jooste, and G. Joubert. 2003. in Preschool Children with Special Reference to Preventability." Annals "Prevalence of Goitre and Urinary Iodine Status of Primary-School of Tropical Paediatrics 16 (2): 141­48. Children in Lesotho." Bulletin of the World Health Organization 81 (1): Molyneux, M. E., T. E. Taylor, J. J. Wirima, and A. Borgstein. 1989. "Clinical 28­34. Features and Prognostic Indicators in Paediatric Cerebral Malaria: A Seely, D. R., S. S. Gloyd, A. D. Wright, and S. J. Norton. 1995. "Hearing Loss Study of 131 Comatose Malawian Children." Quarterly Journal of Prevalence and Risk Factors among Sierra Leonean Children." Archives Medicine 71 (265): 441­59. of Otolaryngology and Head and Neck Surgery 121 (8): 853­58. Msellati, P., P. Lepage, D. G. Hitimana, C. Van Goethem, P. Van de Perre, Sharma, A., P. S. Dhatt, J. C. Lall, H. Singh, H. L. Gupta, and R. N. Sallan. and F. Dabis. 1993. "Neurodevelopmental Testing of Children Born 1976. "Neonatal Tetanus: A Developmental Follow-Up Study." Indian to Human Immunodeficiency Virus Type 1 Seropositive and Pediatrics 13 (1): 51­54. Seronegative Mothers: A Prospective Cohort Study in Kigali, Rwanda." Shrout, P. E., and S. C. Newman. 1989. "Design of Two-Phase Prevalence Pediatrics 92 (6): 843­48. Surveys of Rare Disorders." Biometrics 45 (2): 549­55. Mueller, B. 1994. "Hematological Problems and Their Management in Silver, J., C. E. Gilbert, P. Spoerer, and A. Foster. 1995. "Low Vision in East Children with HIV Infection." In Paediatric AIDS: The Challenge of African Blind School Students: Need for Optical Low Vision Services." HIV Infection in Infants, Children and Adolescents, ed. Phillip Pizzo and British Journal of Ophthalmology 79 (9): 814­20. Catherine Wilfert. Baltimore: Williams and Wilkins. Smith, A., and J. Hatcher. 1992. "Preventing Deafness in Africa's Children." Murray, C. J. 1994. "Quantifying the Burden of Disease: The Technical African Health 15 (1): 33­55. Basis for Disability-Adjusted Life Years." Bulletin of the World Health Organization 72 (3): 429­45. Smith, A. W., J. Hatcher, I. J. McKenzie, S. Thompson, I. Bal, I. Macharia, and P. Mugwe. 1996. "Randomised Controlled Trial of Treatment of Natale, J. E., J. G. Joseph, R. Bergen, R. D. Thulasiraj, and L. Rahmathullah. Chronic Suppurative Otitis Media in Kenyan Schoolchildren." Lancet 1992. "Prevalence of Childhood Disability in a Southern Indian City: 348 (9035): 1128­33. Independent Effect of Small Differences in Social Status." International Journal of Epidemiology 21 (2): 367­72. Snow, R. W., M. Craig, U. Deichmann, and K. Marsh. 1999. "Estimating Mortality, Morbidity and Disability Due to Malaria among Africa's Nicoll, A., I. Timaeus, R. M. Kigadye, G. Walraven, and J. Killewo. 1994. Non-Pregnant Population." Bulletin of the World Health Organization "The Impact of HIV-1 Infection on Mortality in Children Under 77 (8): 624­40. 5 Years of Age in Sub-Saharan Africa: A Demographic and Epidemiologic Analysis." AIDS 8 (7): 995­1005. Sommer, A., and K. P. West. 1996. Vitamin A Deficiency: Health, Survival and Vision. New York: Oxford University Press. Nwosu, S. N. 1998. "Ocular Problems of Young Adults in Rural Nigeria." Stanley, F. J., and L. Watson. 1992. "Trends in Perinatal Mortality and International Ophthalmology 22 (5): 259­63. Cerebral Palsy in Western Australia, 1967 to 1985." British Medical Obama, M. T., L. Dongmo, C. Nkemayim, J. Mbede, and P. Hagbe. 1994. Journal 304 (6843): 1658­63. "Stroke in Children in Yaounde, Cameroon." Indian Pediatrics 31 (7): Steen, R. G., X. Xiong, R. K. Mulhem, J. W. Langston, and W. C. Wang. 791­95. 1999. "Subtle Brain Abnormalities in Children with Sickle Cell O'Shea, T. M. 2002. "Cerebral Palsy in Very Preterm Infants: New Disease: Relationship to Blood Hematocrit." Annals of Neurology 45 Epidemiological Insights." Mental Retardation and Developmental (3): 279­86. Disabilities Research Reviews 8 (3): 135­45. Stein, Z., L. Belmont, and M. Durkin. 1987. "Mild Mental Retardation O'Shea, T. M., K. L. Klinepeter, and R. G. Dillard. 1998. "Prenatal Events and Severe Mental Retardation Compared: Experiences in Eight Less and the Risk of Cerebral Palsy in Very Low Birth Weight Infants." Developed Countries." Upsala Journal of Medical Sciences (Suppl. 44): American Journal of Epidemiology 147 (4): 362­69. 89­96. O'Sullivan, J., C. Gilbert, and A. Foster. 1997. "The Causes of Childhood Swart, S. M., R. Lemmer, J. N. Parbhoo, and C. A. Prescott. 1995. "A Survey Blindness in South Africa." South African Medical Journal 87 (12): of Ear and Hearing Disorders amongst a Representative Sample of 1691­95. Grade 1 Schoolchildren in Swaziland." International Journal of Pediatric Peltola, H. 2001. "Burden of Meningitis and Other Severe Bacterial Otorhinolaryngology 32 (1): 23­34. Infections of Children in Africa: Implications for Prevention." Clinical Tall, F., A. Elola, F. Vincent-Ballereau, and T. Prazuck. 1994. "Anti- Infectious Diseases 32 (1): 64­75. Haemophilus Influenzae B (Hib) Natural Immunity in Children in Robertson, S. 2000. "Background Document: Data on Rubella Vaccination Burkina Faso." Archives of Pediatrics 1 (2): 143­46. Schedules." Report of a meeting on preventing CRS, WHO, Geneva, Tamrat, G., Y. Kebede, S. Alemu, and J. Moore. 2001. "The Prevalence and January 12­14. Characteristics of Physical and Sensory Disabilities in Northern Robertson, S. E., F. T. Cutts, R. Samuel, and J. L. Diaz-Ortega. 1997. Ethiopia." Disability and Rehabilitation 23 (17): 799­804. "Control of Rubella and Congenital Rubella Syndrome (CRS) in Thorburn, M. J., P. Desai, and L. L. Davidson. 1992. "Categories, Classes Developing Countries. Part 2. Vaccination against Rubella." Bulletin of and Criteria in Childhood Disability--Experience from a Survey in the World Health Organization 75 (1): 69­80. Jamaica." Disability and Rehabilitation 14 (3): 122­32. Roeleveld, N., G. A. Zielhuis, and F. Gabreels. 1997. "The Prevalence of Topp, M., P. Uldall, and J. Langhoff-Roos. 1997. "Trend in Cerebral Palsy Mental Retardation: A Critical Review of Recent Literature." Birth Prevalence in Eastern Denmark: Birth Period 1979­86." Developmental Medicine and Child Neurology 39 (2): 125­32. Paediatric and Perinatal Epidemiology 11 (4): 451­60. SADHS (South African Demographic and Health Survey). 1998. Child Trad, P. V., M. Kentros, G. E. Solomon, and E. R. Greenblatt. 1994. Health. Tygerberg: South African Medical Research Council and Macro "Assessment and Psychotherapeutic Intervention for an HIV-Infected International. Preschool Child." Journal of the American Academy of Child and Schneider, M. 1997. We Also Count. The Extent of Moderate to Severe Adolescent Psychiatry 33 (9): 1338­45. Reported Disabilities and the Nature of the Disability Experience in UNICEF (United Nations Children's Fund). 1999. Evaluation of Multiple South Africa. Pretoria: Community Agency for Social Enquiry. Indicator Cluster Surveys. New York: UNICEF. Schulz, K. F., W. Cates Jr., and P. R. O'Mara. 1987. "Pregnancy Loss, Infant UNICEF (United Nations Children's Fund) and Tulane University. 1998. Death, and Suffering: Legacy of Syphilis and Gonorrhoea in Africa." "Progress in Controlling Vitamin A Deficiency." Micronutrient Genitourinary Medicine 63 (5): 320­25. Initiative. Ottawa: UNICEF. 146 | Geoff Solarsh and Karen J. Hofman United Nations Population Division. 1998. World Contraceptive Use. New ------. 1995b. "Prevention of Hearing Impairment: Resolution of the 8th York: United Nations. World Health Assembly, 48.9." WHO, Geneva. ------. 1999. The Demographic Impact of HIV/AIDS. Report of the ------. 1996. Indicators for Assessing Vitamin A Deficiency and Their Technical Meeting. New York: United Nations. Application in Monitoring and Evaluating Intervention Programmes. WHO/NUT/96.10. Geneva: WHO. van Hensbroek, M., A. Palmer, S. Jaffar, G. Schneider, and D. Kwiatkowski. 1997. "Residual Neurologic Sequelae After Childhood Cerebral ------. 1997. Neonatal Tetanus--Progress towards the Global Elimination Malaria." Journal of Pediatrics 131 (1 Part 1): 125­29. of Neonatal Tetanus, 1990­1997. Geneva: WHO. van Stuijvenberg, M. E., J. D. Kvalsvig, M. Faber, M. Kruger, D. G. Kenoyer, ------. 1998. Global Database on Anaemia. Geneva, WHO. and A. J. Banade. 1999. "Effect of Iron-, Iodine-, and Beta-Carotene- ------. 1999. "Approaches for Control of Congenital Disorders and Fortified Biscuits on the Micronutrient Status of Primary School Disability in Primary Health Care." Report of a WHO Meeting Held in Children: A Randomized Controlled Trial." American Journal of Cairo December 6­8. WHO, Geneva. Clinical Nutrition 69 (3): 497­503. ------. 2000. Malnutrition, the Global Picture. Geneva: WHO. Velema, J. P., E. M. Alihonou, T. Gandaho, and F. H. Hounye. 1991. ------. 2001. International Classification of Functioning, Disability and "Childhood Mortality among Users and Non-Users Of Primary Health Health. Geneva: WHO. Care in a Rural West African Community." International Journal of WHO, UNESCO (United Nations Educational, Scientific, and Cultural Epidemiology 20 (2): 474­79. Organization), and ILO (International Labour Office). 1994. Venter, P. A., A. L. Christianson, C. M. Hutamo, M. P. Makhura, and G. S. "Community-Based Rehabilitation for and with People with Gericke. 1995. "Congenital Anomalies in Rural Black South African Disabilities." Joint Position Paper, WHO, Geneva. Neonates--A Silent Epidemic?" South African Medical Journal 85 (1): WHO, UNICEF (United Nations Children's Fund), and ICCIDD 15­20. (International Council for Control of Iodine Deficiency Disorders). Viljoen, D. 1999. "Fetal Alcohol Syndrome." South African Medical Journal 1999. Progress towards the Elimination of Iodine Deficiency Disorders. 89 (9): 958­60. WHO/NHD/99.4. Geneva: WHO. Walker, N., B. Schwartlander, and J. Bryce. 2002. "Meeting International Windham, G. C., and L. D. Edmonds. 1982. "Current Trends in the Goals in Child Survival and HIV/AIDS." Lancet 360 (9329): 284­89. Incidence of Neural Tube Defects." Pediatrics 70 (3): 333­37. WHO (World Health Organization). 1980. International Classification of Winick, M., and A. Noble. 1966. "Cellular Response in Rats during Impairments, Disabilities and Handicaps. Geneva: WHO. Malnutrition at Various Ages." Journal of Nutrition 89 (3): 300­06. Wirz, S., and M. Thomas. 2002. "Evaluation of Community-Based ------. 1992. Prevention of Childhood Blindness. Geneva: WHO. Rehabilitation Programmes: A Search for Appropriate Indicators." ------. 1994. Guidelines for Control of Haemoglobin Disorders. International Journal of Rehabilitation Research 25 (3): 163­71. WHO/HDP/HB/GL/94.1. Geneva: WHO. Zaman, S. S., N. Z. Khan, S. Islam, S. Banu, S. Dixit, P. Shrout, and ------. 1995a. "Global Prevalence of Vitamin A Deficiency. Micronu- M. Durkin. 1990. "Validity of the `Ten Questions' for Screening Serious trient Deficiency Information System." WHO/NUT/95.3. Working Childhood Disability: Results from Urban Bangladesh." International Paper 2, WHO, Geneva. Journal of Epidemiology 19 (3): 613­20. Developmental Disabilities | 147 Chapter 11 Acute Respiratory Infections Shabir A. Madhi and Keith P. Klugman Acute respiratory infections (ARI), particularly lower respi- pollution and overcrowding in households. Reducing these ratory tract infections (LRTI), are the leading cause of death risk factors may take some time, but recent advances in among children under five years of age and are estimated to medical science hold promise in regard to preventing mor- be responsible for between 1.9 million and 2.2 million child- bidity and possibly mortality from the most common per- hood deaths globally. Forty-two percent of these ARI- ceived causes of severe LRTI--those due to bacteria, such as associated deaths occur in Africa (Williams 2002). Despite Haemophilus influenzae type b (Hib) and Streptococcus its importance in regard to morbidity as well as childhood pneumoniae (Cutts et al. 2005; Klugman et al. 2003; mortality, the epidemiology and pathogenesis of LRTI, par- Mulholland et al. 1997). Despite these advances, the chal- ticularly in Africa, remains understudied and consequently lenge remains to address the inequity in health care accessi- underappreciated. Although structured management pro- bility and affordability of new-generation vaccines that have grams coordinated by the World Health Organization been found to be effective in preventing disease caused by (WHO) made some strides during the 1980s and early these bacteria in developed and developing countries. 1990s toward reducing childhood mortality from LRTI Furthermore, a priority for most countries in Sub-Saharan (Sazawal and Black 2003; WHO 1990), the HIV epidemic in Africa in dealing with the reversal of gains that occurred many countries of Sub-Saharan Africa has reversed many of during the late 1980s and early 1990s is to work to prevent these gains (Walker, Schwartlander, and Bryce 2002). The the transmission of HIV to children through effective HIV reduction of morbidity and mortality in Sub-Saharan Africa mother-to-child transmission prevention programs. requires a multifaceted approach that includes addressing risk factors associated with increased susceptibility to LRTI among children. These factors include lack of access to basic THE EPIDEMIOLOGY OF ARI AND LRTI amenities, such as adequate housing, electricity, and run- ning tap water. Availability of these amenities would help Despite the recognition of LRTI as the leading cause of to reduce exposure to such risk factors as indoor smoke childhood mortality in Sub-Saharan Africa (Williams 2002), 149 data on the epidemiology of LRTI in these countries are despite a similar incidence rate of ARI (12.7­16.1 per sparse. The most recent community- and hospital-based 100 child-weeks) in most studies, except the study from longitudinal studies aimed at measuring the burden of LRTI Thailand (Selwyn 1990), suggest widely varying risk factors, in African countries were conducted during the mid-1980s which may have predisposed the children to progression (Bale 1990; Selwyn 1990), prior to the HIV epidemic that from upper respiratory tract infections to LRTI. Another has engulfed most of these countries (Asamoah-Odel et al. possibility is that there remain significant differences in the 2003). These studies were supported as part of an effort by study methodology, including the implementation of differ- the Board of Science and Technology for International ent study definitions between the studies. This underlines Development (BOSTID) of the U.S. National Research the importance of recent initiatives by the WHO to stan- Council to define the burden of LRTI among developing dardize the definition of pneumonia so that comparisons countries (Bale 1990; Selwyn 1990). Based on an estimate of may be made between bacterial conjugate vaccines efficacy ARI, it was concluded that approximately 4 million deaths trials on the burden of LRTI in developing countries (Black from ARI occurred annually (Leowski 1986). In addition to et al. 2002; Cherian et al. 2005; Cutts et al. 2005; Klugman et the African studies sponsored by BOSTID, namely, rural al. 2003; Mulholland et al. 1997). Kenya and urban Nigeria (Oyejide and Osinusi 1990; Wafula The study in Kenya suggested that, in addition to the high et al. 1990), the only other areas from which there are reli- burden of ARI and LRTI among children, on average, ARI able estimates of the burden of LRTI are rural areas in The accounted for their illness 21.7 percent of the time and LRTI Gambia and Ghana (Afari 1991; Campbell et al. 1989). 0.4 percent of the time (Selwyn 1990). Studies from other The literature published between 1966 and 2000 on the non-African developing countries suggest that the burden incidence of LRTI (including that in Sub-Saharan Africa) of disease may be even greater elsewhere, with children has recently been reviewed by Rudan and colleagues, who being affected for 40 to 60 percent and 14 percent of the year estimated the median incidence of LRTI in developing with ARI and LRTI, respectively (Black et al. 1982; Selwyn countries at 44 episodes per 100 child-years, equal 1990). The prolonged morbidity associated with ARI, in to approximately 150.7 million new cases each year, 7 to addition to being associated with mortality, may also affect 13 percent of which were severe enough to warrant hospi- the physical, emotional, and intellectual development of the talization (Rudan et al. 2004). They estimate that there were child. This situation is further compounded by the finding approximately 33.7 million cases of LRTI annually in that the rates of ARI are greatest during the initial years of African children, with an incidence rate of 31 to 33 per 100 life (about 7.5 to 8 episodes per year among infants versus child-years, which was higher than the overall average for all about 4 to 6 episodes per year among children 48 to 59 developing countries (23 per 100 child-years) (Rudan et al. months of age; Selwyn 1990). Furthermore, the proportion 2004). These estimates were, however, exclusive of the of time that children are affected by an ARI ranged from impact that the HIV epidemic may have had on the inci- about one-third of the year among infants to less than dence of LRTI. Rudan and colleagues (2004) also found that 20 percent of the time among older children (Oyejide and the burden of LRTI, albeit from studies outside of Africa, Osinusi 1990; Wafula et al. 1990). The study from Nigeria was greatest among children less than one year old, and, rel- also documented the benefits of measles immunization on ative to an incidence of 1.0 in the first year of life, the mean the incidence of LRTI, showing that children vaccinated for ratios for the subsequent four years of life were 0.58 (year 2), measles had a 2.8-fold reduction in the incidence of LRTI 0.48 (year 3), 0.31 (year 4), and 0.19 (year 5). (Oyejide and Osinusi 1990). Although the few published studies allow an analysis of The only longitudinal burden-of-disease study in Africa the burden of LRTI mainly during the mid-1980s and early to have defined the incidence of radiologically confirmed 1990s, a major pitfall in the studies was the differences in pneumonia was performed in The Gambia (Campbell et al. the methodology, not only of diagnostic tests, but also of the 1989). In that study the incidence of LRTI was 45 per 100 clinical definitions used in the studies, all of which may child-years, and 35.5 percent of episodes were associated have a bearing on the measured incidence rates as well as the with radiologically confirmed pneumonia. Another cross- described etiology of LRTI. Although the studies sponsored sectional study done in the Central African Republic also by BOSTID aimed at reducing the differences in methodol- found that, although radiologically confirmed pneumonia ogy between studies, the wide range of the incidence of was present in only 41 percent of children hospitalized for LRTI (0.4­8.1 per 100 child-weeks) between the studies, LRTI, the presence on chest radiograph of the involvement 150 | Shabir A. Madhi and Keith P. Klugman of three or more lobes was associated with a 4.6-fold greater estimated 4.5 percent of children in the study population risk of death than that of a single lobe (Demers et al. 2000). being HIV infected, these children accounted for 45 percent The overall mortality rate from LRTI among children in of all LRTI-associated hospitalizations (Madhi, Petersen, developing countries based on the longitudinal studies Madhi, Khoolsal, et al. 2000). reviewed by Rudan and colleagues (2004) ranged from 6.6 Although there are no longitudinal studies from to 14.1 percent. Sub-Saharan Africa, including surveillance at the The high burden of LRTI among developing countries is community­health care level, of the incidence of LRTI partly related to the increased exposure of children in these among HIV-infected children, data from developed coun- countries to those poverty-linked risk factors that predis- tries illustrate the heightened burden of LRTI among HIV- pose them to developing LRTI. In a multivariate analysis of infected children. In a recent review of 3,331 children data from various studies, Rudan and colleagues (forthcom- infected with HIV-1 who participated in several different ing) identified the risk factors most consistently associated clinical trials, the incidence rate of any clinically or radio- with LRTI as malnutrition (relative risk [RR] 1.8), low logically diagnosed pneumonia was estimated to be 11.1 birthweight (less than 2,500 grams at birth, RR 1.4), lack of (95 percent confidence interval [CI], 10.3­12.0) per 100 exclusive breastfeeding in the first four months of life child-years (Dankner, Lindsey, and Levin 2001). The trials (RR 1.3), overcrowding in the household (more than five included children who were on one to three antiretroviral people in the household, RR 1.2), and lack of measles drugs, excluding protease inhibitors. The mean age of chil- immunization (RR 0.7). It was estimated that these risk fac- dren in this study was 39 months; the oldest subject was tors were prevalent among 27.4 percent, 14.7 percent, 54.1 20.9 years of age (Dankner, Lindsey, and Levin 2001). In an percent, 80.3 percent, and 52.5 percent, respectively, of the earlier cohort study evaluating the efficacy of intravenous world's 523 million children age zero to four years. immunoglobulin prophylaxis, performed among HIV- Additional poverty-related factors that predispose to LRTI infected children in the United States, the incidence of LRTI include domestic smoke and air pollution (Anderson 1978; was estimated to be 24 per 100 child-years (Mofenson et al. Sofoluwe 1968). Consequently, it can be extrapolated that a 1998). This was sevenfold greater than historical incidence significant burden of LRTI among children in the develop- rates of 3 to 4.2 per 100 child-years observed among pre- ing world is avoidable by addressing factors associated with school children in the United States and Finland prior to the poverty and lower socioeconomic development. HIV-1 epidemic (Foy et al. 1973; Murphy et al. 1981). Mofenson and colleagues (1998) described the increased burden of LRTI among HIV-infected children but underes- THE IMPACT OF THE HIV EPIDEMIC timated the true burden of disease in young children, ON THE EPIDEMIOLOGY OF LRTI since the mean age of children recruited into the study (40 months) was greater than the ages (less than 18 months A further limitation of the published longitudinal studies is of age) when HIV-infected and HIV-uninfected children are that they were conducted at an early stage or even prior to most susceptible to developing and dying of pneumonia the HIV epidemic. This is particularly pertinent when con- (Chintu et al. 1995; Langston et al. 2001; Mofenson et al. sidering that in some centers more than 80 percent of deaths 1998; Pillay et al. 2001; Selwyn 1990). During the course of from LRTI occur among children infected with the human the study by Mofenson and colleagues (1998), pneumonia immunodeficiency virus (HIV), despite only 4 to 6 percent was second (12 percent) only to upper respiratory tract of children in the general pediatric population being so infections as a cause of bacterial infections among HIV- infected (Madhi, Petersen, Madhi, Khoolsal, et al. 2000; Zwi, infected children. Thirty-seven percent of the children with Pettifor, and Soderlund 1999). Although a few studies have acute pneumonia had multiple episodes of pneumonia, looked at differences in the etiology of LRTI between HIV- which was also a risk factor for mortality (odds ratio [OR], infected and HIV-uninfected children, only one published 2.1; 95 percent CI, 1.3­3.4). study has described the impact that the HIV epidemic has Data from South Africa indicated that the incidence had on the relative burden of LRTI among children in of hospitalization for LRTI among nearly 20,000 placebo Africa. This study was limited to hospital-based surveil- recipients (who all received Hib conjugate vaccine) partici- lance and underestimates the overall burden of LRTI. pating in a pneumococcal conjugate vaccine trial, of whom Nevertheless, the study emphasized that despite only an an estimated 6.04 percent were infected with HIV-1, was Acute Respiratory Infections | 151 approximately 6.6 times greater among HIV-infected chil- however, have possibly been reversed due to the impact that dren (16.7 cases per 100 child-years) than HIV-uninfected the HIV epidemic has had on childhood mortality in Sub- children (2.6 cases per 100 child-years) (Madhi et al. 2005). Saharan Africa since 1990 (Walker, Schwartlander, and Although the incidence rates observed for HIV-infected chil- Bryce 2002). In addition to the increased predisposition of dren were less than the rate observed by Mofenson and col- HIV-infected children to bacterial- and viral-associated leagues (1998) among HIV-infected children in the United LRTI, HIV-infected children (13.1 percent) have also been States (24 per 100 child-years), the data from South Africa found to have a 6.5 times greater (95 percent CI, 3.5­12.1) included only children who were hospitalized for LRTI as case-fatality rate than HIV-uninfected children (2.3 per- opposed to children who were hospitalized for LRTI and cent) even in a country with relatively good resources, such ARI and those who were outpatients in the United States. as South Africa (Madhi, Petersen, Madhi, Khoolsal, et al. Although differences in the age of children are also impor- 2000). Although the case-fatality rate for children with tant, these data may illustrate the increased burden of LRTI pneumococcal bacteremic pneumonia was similar among among African HIV-infected children compared with HIV- those infected with HIV (18 percent) compared with those infected children in industrial countries. uninfected (11 percent, P 0.18), the overall higher case- Williams (2002) recently reviewed the estimates of child- fatality rate for LRTI among HIV-infected children may be hood deaths due to ARI mainly based on published data explained by the poor outcome of African HIV-infected prior to the current HIV epidemic. It was estimated that children who have Pneumocystis carinii pneumonia (PCP) globally ARI causes 1.9 million (95 percent CI, 1.6 million to (case-fatality rate 20 to 65 percent) (Graham et al. 2000; 2.2 million) deaths annually of children under five, 794,000 Madhi et al. 2002). (40 percent) of which occur in Africa (Williams 2002). The proportion of deaths attributable to ARI was found to be related to the under-five childhood mortality rate of the THE ETIOLOGY OF LRTI countries. In countries where the under-five childhood mortality rate is 50 per 1,000, as in most of Sub-Saharan Unfortunately, the sensitivity of current diagnostic tools in Africa (UNICEF 1996, pp. 90­98), ARI is estimated to be defining the etiology of LRTI is woefully suboptimal, partic- responsible for 23 percent of deaths, compared with only ularly for diagnosing bacterial infections, even in the coun- 15 percent of deaths in countries where the under-five mor- tries with the most resources, such as Finland and the tality rate is 10 per 1,000 per year (Williams 2002). The United States (Heiskanen-Kosma et al. 1998; Wubbel et al. lower contribution of ARI as a cause of childhood deaths as 1999). Among industrial countries, where the common res- under-five mortality rates improve was further character- piratory viruses, particularly respiratory syncytial virus, are ized by trends in under-five mortality among black South the dominant (40 to 50 percent of episodes) pathogens Africans that were observed between the 1960s and 1980s identified among children with LRTI, no etiological agent (Von Schirnding, Yach, and Klein 1991). In the indigenous was identified among 30 to 40 percent of children with South African population of black people from 1968 to LRTI, despite the use of an array of advanced microbiologi- 1973, a corrected estimate of 22.5 percent of the under-five cal methods, including some that had not been validated mortality of 40 per 1,000 was attributed to ARI, compared against "gold standards" (Heiskanen-Kosma et al. 1998; with 1980­85, when, with an under-five mortality rate of Wubbel et al. 1999). 17.3 per 1,000, the proportion of deaths due to ARI The case-management strategy of the WHO is based on decreased to 17.7 percent (Von Schirnding, Yach, and Klein the premise that bacteria are the leading cause of LRTI, 1991). Similarly, ARI was linked to a high proportion (21 to especially as a cause of severe illness, among children in 24 percent) of deaths in other African countries, where the developing countries (WHO 1990). Nevertheless, most of under-five mortality rate ranged between 35 and 63 per the etiological studies performed in developing countries 1,000 children during the 1980s (de Francisco et al. 1993; suggest that respiratory viruses are as important a cause of Fantahun 1998; Greenwood et al. 1987; Mtango and LRTI among these children as among those in developed Neuvians 1986; Williams 2002). countries (Forgie et al. 1991a, 1991b; Heiskanen-Kosma Much of the improvement in under-five mortality as well et al. 1998; Selwyn 1990; Wubbel et al. 1999). as the reduction in the proportion of deaths due to ARI that The common respiratory viruses isolated among African has been observed in other Sub-Saharan Africa countries, infants hospitalized for LRTI in The Gambia include 152 | Shabir A. Madhi and Keith P. Klugman respiratory syncytial virus (37 percent), adenovirus (5 per- Advances in preventing LRTI caused by S. pneumoniae cent), parainfluenza virus (3 percent), rhinovirus (6 per- and H. influenzae type b have refocused attention to meas- cent), and influenza virus (1 percent) (Forgie et al. 1991a, uring the burden of disease due to these pathogens in order 1991b). Furthermore, more recently, a newly discovered res- to define the burden of LRTI that may be prevented through piratory virus named metapneumovirus (van den Hoogen vaccination of children with the bacterial conjugate vaccines et al. 2001) has also been identified as a cause of severe LRTI (Black et al. 2002; Klugman et al. 2003; Mulholland et al. among 10 percent of HIV-uninfected and 3 percent of HIV- 1997). Although fine needle aspiration of the lung is the infected infants (Madhi et al. 2003). Bacteria were, however, gold standard for identifying bacterial infections of the lung, proportionally as common and were identified in 30 percent it only has a sensitivity of 70 percent and is limited to being of the infants, albeit using unvalidated diagnostic methods, performed mainly on children with clearly defined alveolar with the dominant identified pathogens being S. pneumoniae consolidation, accessible to aspiration (Vuori-Holopainen (20 percent) and H. influenzae (11 percent) (Forgie et al. and Peltola 2001). Because of the empirical management of 1991a). Mixed bacterial-viral infections were identified severe LRTI with antibiotics, the use of fine needle aspirates among 15 percent of infants. Although 73 percent of infants of the lung has waned (Vuori-Holopainen and Peltola 2001) had radiological evidence of pneumonia, only 28 percent but has recently received more attention as a diagnostic had evidence of a lobar pneumonia( Forgie et al. 1991a). approach in Finland (Vuori-Holopainen et al. 2002). Among older children (one to four years of age) hospitalized The majority of studies that have used fine needle aspi- with LRTI in The Gambia, identification of bacteria was rates as a potential diagnostic modality have been performed relatively more common (75 percent) than identification in Africa (Vuori-Holopainen and Peltola 2001). The per- of common respiratory viruses (40 percent) (Forgie et al. formance of lung aspirates is, however, associated with some 1991b). Although S. pneumoniae (60.9 percent) and risk (less than 2 percent), such as that of pneumothoraces H. influenzae (13.8 percent) remained the most dominant and hemoptysis, which may preclude its use in clinical prac- bacterial isolates, other bacteria identified included tice as well as in large-scale epidemiological studies (Vuori- Staphylococcus aureus and Moraxella catarrhalis (6.2 percent Holopainen and Peltola 2001). Furthermore, the yield from each) (Forgie et al. 1991b). These studies, similar to others lung aspirates may be operator dependent and could also be from developing countries, including those that were spon- influenced by other factors, such as preceding exposure to sored by BOSTID, underpin the importance of respiratory antibiotics, age of the study participants, and the extent of syncytial virus as well as S. pneumoniae, H. influenzae, and lung infiltrate. This may, at least in part, explain why the yield viruses as the etiology of LRTI among children in develop- ranged from 17 percent to 100 percent in various studies. On ing countries (Selwyn 1990). average, a bacterial pathogen from lung aspirates was isolated The spectrum of pathogens and proportional representa- in 52 percent (17 to 77 percent) of cases, compared with a tion thereof appeared to be similar between children with blood isolation rate of 25 percent (2 to 45 percent) in studies moderate and severe LRTI in the BOSTID-sponsored stud- in which both procedures were done concurrently (Adegbola ies (Selwyn 1990). The clinical relevance of distinguishing et al. 1994; Garcia de Olarte et al. 1971; George, Bai, and between respiratory viral and bacterial infections has, how- Cherian 1996; Rapkin 1975; Silverman et al. 1977; Vuori- ever, gained renewed interest with the recent observation Holopainen and Peltola 2001; Wall et al. 1986). that pneumococcal coinfections are prevalent in at least 30 Table 11.1 illustrates those studies that included only to 40 percent of children with documented respiratory viral children with pneumonia who had not received antibiotics infections (Madhi, Klugman, and Pneumococcal Vaccine prior to having had a fine needle aspirate performed. Except Trialist Group 2004). for a small study from South Africa, all the studies con- Other factors that may affect the spectrum of bacteria that firmed S. pneumoniae as the leading isolate among children causes pneumonia include environmental factors and nutri- with pneumonia, although there was a wide range of posi- tional status (Adegbola et al. 1994; O'Dempsey et al. 1994). tive pneumococcal isolates (18 to 51 percent) among other Gram-negative bacteria that include Salmonella and coliforms studies performed in developing countries. A summary of were found to be more common than S. pneumoniae and Hib lung aspirate studies performed in Africa has, however, sug- as causes of LRTI during those rainy months of the year when gested that S. aureus may also be an important cause of LRTI malaria transmission is at its greatest in a malaria-endemic (20 percent of children) (Vuori-Holopainen et al. 2002; area in The Gambia (O'Dempsey et al. 1994). Mimica et al. 1971). Acute Respiratory Infections | 153 Table 11.1 Results of Lung Aspirations of Children Who Had Not Received Antibiotics and Not Mentioned Underlying Illness No. of Percentage testing Other Country Age children positivea Pneumococcus H. influenzae S. aureus E. coli bacteria USA 2 mos.­15 yrs. 27 22 50 (11) 0 (0) 0 (0) 0 (0) 50 (11) South Africa 2 mos.­9 yrs. 29 17 20 (3) 40 (6) 0 (0) 0 (0) 40 (3) Nigeria 4 mos.­ 8 yrs. 88 79 64 (51) 14 (11) 14 (11) 0 (0) 31 (39) Zimbabwe 2 mos.­11 yrs. 40 33 54 (18) 23 (8) 31 (10) 0 (0) 0 (0) Papua New Guinea 10 yrs. 18 44 88 (39) 13 (6) 0 (0) 0 (0) 0 (0) Sources: Data from Rapkin 1975 (United States); Prinsloo and Cicoria 1974 (South Africa); Silverman et al. (Nigeria) 1977; Ikeogu (Zimbabwe) 1998; and Riley et al. (Papua New Guinea) 1983. Note: Numbers in parentheses are percentages. a. Value indicated in the cells under bacteria is the proportion of all positive isolates and in parenthesis are the values as a percentage of all tests performed. More than one bacterium may have been isolated in some children. Table 11.2 Estimated Incidence of Organism-Specific, Bacteremic, Community-Acquired Severe LRTI in HIV-1-Infected and HIV-1-Uninfected Children Age 2 to 24 Months HIV-1-infected children HIV-1-uninfected children Organism (per 100,000) (per 100,000) Relative risk; 95% CI; P value Streptococcus pneumoniae 1,233 29 42.9; 20.7­90.2; P 10 5 Haemophilus influenzae type b 569 27 21.4; 9.4­48.4; P 10 5 Staphylococcus aureus 337 3 49.0; 15.4­156.0; P 10 5 Escherichia coli 474 10 97.9; 11.4­838.2; P 10 5 Salmonella species 95 7 13.4; 2.2­78.1; P 0.02 Mycobacterium tuberculosis 1,470 65 22.5; 13.2­37.6; P 10 5 Source: Madhi, Petersen, Madhi, Khoolsal, et al. 2000. Additionally, investigators in Chile and India reported Data regarding the relative risk of bacterial pneumonia as that S. aureus was the dominant isolate (25 to 27 percent), well as common respiratory viral-associated LRTI among whereas S. pneumoniae was isolated in only 2 to 5 percent of African HIV-infected children are few (Madhi, Petersen, children, among whom 50 to 70 percent had received antibi- Madhi, Khoolsal, et al. 2000; Madhi, Schoub, et al. 2000). otics prior to the performance of the lung aspirate (Mimica Although many studies have investigated the etiology of LRTI et al. 1971). In another report among children who were among African HIV-infected children (Graham et al. 2000; receiving antibiotics (77 percent) prior to the lung aspirate, Nathoo et al. 1996; Zar et al. 2001), only limited data exist S. pneumoniae was isolated more frequently than S. aureus regarding the pathogen-specific relative risk of LRTI among (11 percent as opposed to 6 percent, respectively) (Prakash HIV-infected compared with uninfected children. Table 11.2 et al. 1996). Silverman and colleagues (1977) showed that illustrates the heightened burden of pathogen-specific bac- although the yield of bacteria in Nigerian children with an teremic pneumonia among HIV-infected as compared with ill-defined infiltrate (bronchopneumonia) was similar to HIV-uninfected children. Significantly, although a broader that in children with lobar consolidation or empyema, the spectrum of bacteria is implicated in HIV-infected children-- spectrum of isolates differed, with the yield of S. pneumoni- in particular, gram-negative pathogens--S. pneumonia and ae being 53.6 and 15.9 percent, respectively (Silverman et al. Hib remained the dominant causes of pneumonia among 1977). Similarly, Mimica and colleagues (1971) in Chile HIV-infected children in the absence of vaccination with any showed that although a bacterium was isolated in 28 percent of the bacterial conjugate vaccines. In addition to the height- compared with 45 percent of children with lobar consolida- ened burden of bacterial pneumonia, although proportion- tion and bronchopneumonia, respectively, the rates of isola- ately less commonly isolated, the burden of hospitalization tion of S. pneumoniae were much lower among children for common respiratory viral infections was also found to with bronchopneumonia (1 percent versus 24 percent) have increased, albeit less so than for bacterial pneumonia, (Mimica et al. 1971). among HIV-infected children (table 11.3). 154 | Shabir A. Madhi and Keith P. Klugman Table 11.3 Estimated Incidence for Specific Viral-Associated THE IMPACT OF WHO MANAGEMENT STRATEGY Severe LRTI in HIV-1-Infected and HIV-1-Uninfected Children Age 2 to 23 Months Following the realization that LRTI is a major contributor to childhood mortality, the WHO developed a strategy aimed HIV-1-infected HIV-1-uninfected children children Relative risk, at allowing primary health workers to manage LRTI speedily Virus (per 100,000) (per 100,000) 95% CI and effectively. The strategy, which subsequently has been RSV 1,444 309 1.92, 1.29­2.83 incorporated into the integrated management of childhood Influenza A/B 1,268 148 8.03, 5.05­12.76 illness (IMCI) program, is based on the premises that the Parainfluenza 1-3 893 106 8.46, 4.95­10.47 majority of severe LRTI in developing countries is due to Adenovirus 481 32 15.07, 6.62­34.33 bacteria, particularly S. pneumoniae and H. influenzae, and that proper management of the disease needs to include Source: Madhi, Schoub, et al. 2000. consideration of the scarcity of health care facilities in those countries, which also coincidentally happen to be the most burdened by LRTI. A meta-analysis of this strategy has Further compounding the impact of the HIV epidemic shown that it has been most effective when implemented in on the etiology of LRTI and possibly the WHO case- those countries where the infant mortality rate was greater management strategy for its empirical management in than 100 per 1,000 live births (Sazawal and Black 2003). The Sub-Saharan Africa is the dominance of other opportunistic meta-analysis found that the WHO LRTI case-management pathogens. Pneumocystis jiroveci, which is an uncommon strategy reduced LRTI mortality by 42 percent (95 percent cause of LRTI in HIV-uninfected children, except perhaps CI, 22­57 percent), 36 percent (95 percent CI, 20­48 per- those that are malnourished or who have other immuno- cent), and 36 percent (95 percent CI, 20­49 percent) among suppressive underlying illness (Hughes et al. 1974; Pifer neonates, infants between one and twelve months of age, et al. 1978), is commonly (10 to 45 percent of cases) identi- and children one to four years of age, respectively (Sazawal fied among African HIV-infected children with LRTI and Black 2003). The WHO LRTI management strategy has (Graham et al. 2000; Madhi et al. 2002). Although PCP is more recently also been evaluated in a program in Malawi prevalent in various age groups according to one study in (Enarson and Pio 2003; Pio and Enarson 2003). The infant South Africa (Madhi et al. 2002), both the burden of the dis- mortality rate in the study area prior to the intervention ease and its poor outcome have consistently been found to was 138 per 1,000, and approximately 20 percent of preg- be greatest among children less than six months of age. nant women were HIV infected, resulting in approximately A further recent observation has been the underrecog- 4.5 percent of the birth cohort being HIV infected. nized importance of Mycobacterium tuberculosis as a cause Implementation of a strategy aimed at the training of dis- of acute LRTI. Three separate studies from South Africa and trict health care workers was associated with a 52 percent Malawi consistently reported that M. tuberculosis was cul- reduction in the LRTI mortality if the child survived beyond tured from 8 percent of children with acute pneumonia 24 hours of having presented at a health care facility. There (Graham et al. 2000; Madhi, Petersen, Madhi, Khoolsal, et al. was, however, only a modest nonsignificant reduction 2000; Zar et al. 2001). This is even more alarming consider- (15 percent, P 0.05) in the deaths that occurred within ing that most of these studies were performed among 24 hours of attention at a health care center. Overall, the children who were mainly investigated for tuberculosis by case-fatality rate for LRTI decreased from 18 percent at the gastric washing or induced sputum samples--modalities time of the start of the study to 8 percent 26 months later. that have a sensitivity of only 25 to 40 percent for diagnos- These reductions were observed within two years of having ing pulmonary tuberculosis. Consequently, it can be postu- implemented the training of health care workers in the lated that as many as three times more children than management of LRTI in accordance with the WHO/IMCI diagnosed by positive M. tuberculosis culture have acute strategy as well as ensuring adequate access to essential pneumonia caused by this pathogen. Further investigation drugs. The overall success of treating LRTI also improved of these observations are required and, if verified, would from 55 percent at the time of the start of the program in have profound implications for the clinical algorithms used 1999 to 82 percent 26 months later. These results, from one in diagnosing pulmonary tuberculosis as well as the man- of the poorest nations on earth, show that standard inpa- agement strategies for treating LRTI in children. tient case management of severe and very severe pneumonia Acute Respiratory Infections | 155 by trained staff with a regular supply of antibiotics pro- observation. Although the effectiveness of the Hib conjugate duced a striking impact on the number of deaths occurring vaccine against pneumonia has not been evaluated among after 24 hours of hospital admission, even in the adverse HIV-infected children, data from South Africa suggest that conditions of Malawi, where the prevalence of HIV infection the vaccine is less effective among HIV-infected than HIV- is high, malnutrition is rife, the level of maternal literacy is uninfected children, reducing invasive disease by 54 per- low, and an efficient transport system from peripheral to dis- cent and 91 percent, respectively (Madhi et al. 2005). trict hospitals is lacking. Consequently, it can be extrapolated that the Hib conjugate Considering the numerous challenges to expand and vaccine may also be less effective in preventing nonbac- maintain health care provision in resource-poor countries, teremic pneumonia among HIV-infected children than prevention rather than treatment of LRTI assumes even HIV-uninfected children. greater importance. Although some of the risk factors In The Gambia, prior to the introduction of the Hib discussed earlier can be addressed only by a general conjugate vaccine, H. influenzae was identified as the sec- improvement of socioeconomic conditions in these coun- ond most important cause of bacterial pneumonia, with tries, a potentially more immediate measure would be 75 percent of the isolates being of type b. The importance of vaccination. That immunization may affect the incidence of H. influenzae is further highlighted by lung aspirate studies LRTI-associated mortality is probably best indicated by the in children with pneumonia who did not receive antibiotics. success of the measles vaccine in reducing LRTI-associated Here, Hib accounted for an average of 27 percent (range of deaths in Nigeria (Oyejide and Osinusi 1990). In the 14 to 45 percent) of all bacterial isolates. Indeed H. influen- Nigerian study, the under-five mortality rate due to LRTI zae has been shown to be as common as S. pneumoniae was twofold less common among children who had been (average of 27 percent; range of 12 to 88 percent) as a cause immunized with measles vaccine than among children who of pneumonia in some studies (Berman and McIntosh had not received measles vaccine. 1985). The incidence of bacteremic Hib pneumonia in chil- dren less than five years of age ranged between 1 and 7 per 100,000 children in developed countries, such as Finland THE IMPACT OF HIB CONJUGATE VACCINES and Israel (Dagan et al. 1998; Takala et al. 1989), whereas it IN PREVENTING LRTI was estimated to be 370 per 100,000 in developing countries if nonbacteremic cases were included (Peltola 2000). Prior The potential of further preventing LRTI through vaccina- to the introduction of the Hib conjugate vaccine in The tion was shown in The Gambia during the course of evalu- Gambia, it was estimated that Hib resulted in a pneumonia- ating the efficacy of a Hib conjugate vaccine among infants specific mortality rate of 40 per 100,000 in children zero to (Mulholland et al. 1997). In addition to reducing invasive four years old (Greenwood 1992). Hib disease, including Hib meningitis, the vaccine was also Despite Hib conjugate vaccines being available since the shown to reduce the incidence of radiologically confirmed late 1980s, the only Sub-Saharan African country using its pneumonia by 21 percent (reduction of 1.3 cases per 1,000 own resources to introduce the Hib conjugate vaccine into children). The overall reduction in clinically diagnosed its routine childhood immunization program since June pneumonia was, however, more modest, with a non- 1999 was, until very recently, South Africa. The vaccine has significant reduction of 4.4 percent (95 percent CI, 5 to been introduced into other Sub-Saharan African countries 12.9), although the study was not statistically powered to through donor support from the Global Alliance for measure such small differences against clinically diagnosed Vaccines and Immunisation (GAVI). GAVI has committed LRTI. Interestingly, the incidence rate of clinically diagnosed itself to financially supporting the introduction of the Hib LRTI requiring hospitalization in The Gambian study was conjugate vaccine into countries that have a gross domestic only 3.6 per 1,000 children among the placebo group. product of less than US$2,000 per capita for a period of five The findings from The Gambia were subsequently cor- years, following which the governments of those countries roborated by an effectiveness study in Chile, where a similar would be expected to assume the responsibility of continu- reduction in radiologically (21 percent) and clinically diag- ing to finance the use of the Hib conjugate vaccine. Whether nosed (5 percent) pneumonia was observed (Levine et al. these programs are sustainable is currently being debated. 1999). The Hib conjugate vaccine was also found to prevent Even in an African country with relatively good resources, 2.5 cases of hospitalization for LRTI per 1,000 child-years of such as South Africa, the modest price of the vaccine 156 | Shabir A. Madhi and Keith P. Klugman (approximately US$2 per dose) has almost quadrupled the mainly ambulant cases, in the control group in the latter cost of vaccine procurement for the expanded program of study was 55.9 per 1,000 person-years. The reduction in immunization, and its continued use is under debate by pol- pneumonia confirmed by chest radiograph was striking icy makers. The cost of the full primary series of vaccines insofar as previous epidemiological studies have shown that administered during the initial four months of life prior to between 25 and 35 percent of cases of pneumonia in chil- the introduction of the Hib conjugate vaccine was approxi- dren are due to S. pneumoniae (Heiskanen-Kosma et al. mately US$1.5. 1998; Wubbel et al. 1999), indicating that the vaccine is highly efficacious even for this end point for which there was no microbiologic evaluation. THE IMPACT OF S. PNEUMONIAE CONJUGATE A different formulation of the vaccine that included two VACCINES IN PREVENTING PNEUMONIA more serotypes than those included in Prevenar was subse- IN CHILDREN quently evaluated among children in South Africa and The Gambia. These additional serotypes--serotypes 1 and 5-- Despite S. pneumoniae having first been isolated in the account for about 15 percent of IPD in Africa (Hausdorff, 1880s and having been recognized as the leading cause of Siber, and Paradiso 2001; Madhi, Petersen, Madhi, Wasas, pneumonia among children, a successful vaccine aimed at et al. 2000). In regard to the serotype-specific vaccine efficacy preventing pneumococcal pneumonia has remained elusive against IPD as well as radiologically confirmed pneumonia until 2000 (Black et al. 2002; Klugman et al. 2003). Although among HIV-uninfected children, the results from the South Riley and colleagues (1986) showed a benefit for a pneumo- African study were strikingly similar to the findings in chil- coccal polysaccharide vaccine in reducing childhood mor- dren from the United States. Among South African children, tality in Papua New Guinea, other studies failed to show any the nonavalent pneumococcal conjugate vaccine was shown benefit for the use of this vaccine, particularly among chil- to reduce culture-confirmed IPD by 83 percent and radio- dren less than 18 months of age (Broome and Breiman logically confirmed pneumonia, irrespective of microbio- 1991). Following the success of the Hib conjugate vaccine, logical diagnosis, by 20 percent (Klugman et al. 2003). the same technology was used to produce a pneumococcal Impressively, the vaccine was also shown to reduce IPD by serotype-specific polysaccharide-protein conjugate vaccine. 65 percent among HIV-infected children; however, there Those serotypes that were recognized as causing the major- was no significant reduction in first-episode, radiologically ity of invasive disease among children were selected for confirmed pneumonia among these children (13 percent, inclusion into the vaccine. The first and currently the only 95 percent CI, 7­29) (Klugman et al. 2003). Further analy- licensed such vaccine is one that includes polysaccharide sis of this study, however, indicated that when considering from seven of the pneumococcal serotypes (Prevenar), any LRTI irrespective of the chest radiograph features, the based on those serotypes that were most prevalent as a cause vaccine reduced LRTI by 15 percent (95 percent CI, 6­24) of invasive pneumococcal disease (IPD) among children in among HIV-infected children. This reduction in LRTI the United States (Zangwill et al. 1996). Inclusion of these translated into a reduction of 25.7 cases of LRTI per 1,000 seven serotypes was estimated to have the potential of child-years, compared with a case reduction of 2.7 cases preventing 90 percent of IPD among children in the United per 1,000 child-years in HIV-uninfected children (efficacy States (Zangwill et al. 1996). The vaccine was subsequently 17 percent; 95 percent CI, 7­26) (Madhi et al. 2005). proved to be highly efficacious among those children and Interestingly, the reductions in radiologically confirmed was associated with the reduction of vaccine-serotype- alveolar consolidation­associated LRTI were 1 and 9 per specific invasive disease in fully vaccinated children of 1,000 child-years in HIV-uninfected and HIV-infected chil- 97.4 percent (95 percent CI, 82.7­99.9), with an overall dren, respectively (Madhi et al. 2005). These data suggest reduction of invasive pneumococcal disease of 89.1 percent that the sensitivity of radiologically diagnosed pneumonia, (95 percent CI, 73.7­95.8) (Black et al. 2000). More impor- based on WHO criteria, detected only about one-third of tant, however, the vaccine was shown to reduce pneumonia the cases of pneumococcal pneumonia that were prevented associated with any chest radiograph infiltrate by 20.5 per- by vaccination. Studies that aim to define the burden of cent (95 percent CI, 4.4­34.0) and any clinically diagnosed pneumonia preventable by the pneumococcal conjugate pneumonia by 4.3 percent (95 percent CI, 3.5­11.5 per- vaccine among children in other countries of Sub-Saharan cent) (Black et al. 2002). The incidence rate of pneumonia, Africa need to take these factors into consideration when Acute Respiratory Infections | 157 determining the tools to be used in defining LRTI, including assist in introducing the vaccine, it is likely to lag even radiologically confirmed LRTI. Significantly, the data from beyond the 20 years that it has taken for the Hib conjugate the South African study included only children who vaccine to be introduced into Sub-Saharan Africa. required hospitalization for their episode of LRTI. Provisional data from the South African study indicate that Considering that only 7 to 13 percent of children with LRTI the vaccine would be potentially cost-effective if it cost may have been hospitalized, the total burden of LRTI that approximately US$4 per dose, based on health expenditure may be prevented through vaccination may be much costs in South Africa (Ginsberg et al. unpublished data). greater, especially considering that the spectrum of Considering that health care expenditure in other Sub- pathogens that cause LRTI requiring hospitalization were Saharan Africa countries is much lower than in South found to be similar to those that caused LRTI that did not Africa, the price at which the vaccine may be cost-effective require hospitalization (Selwyn 1990). would possibly be even lower. The nonavalent pneumococcal conjugate vaccine also reduced invasive pneumococcal disease (77 percent; 95 per- cent CI, 51­90) and radiologically confirmed pneumonia OTHER POTENTIAL INTERVENTION STRATEGIES (37 percent; 95 percent CI, 25­48) in The Gambia, which is IN REDUCING THE BURDEN OF LRTI much less developed than South Africa. Impressively, in addition to a reduction in cases of pneumonia of 16 per The burden of LRTI may be reduced by as much as 45 per- 1,000 child-years among vaccinees in The Gambia, the non- cent in countries of Sub-Saharan Africa (Madhi, Petersen, avalent pneumococcal conjugate vaccine was also found to Madhi, Wasas, et al. 2000) if effective strategies targeted at reduce all-cause hospitalization by 15 percent (95 percent preventing HIV transmission from mother to children are CI, 7­21) and all-cause mortality by 16 percent (95 percent implemented. Although a number of proven strategies CI, 3­28) (Cutts et al. 2005). These data clearly indicate the could reduce the vertical transmission of HIV to between 2 urgency attendant on the introduction of this new vaccine and 13 percent, lack of infrastructure in Sub-Saharan Africa into developing countries. Modification of the currently results in less than 5 percent of HIV-infected women being licensed product, so as to include pneumococcal serotypes able to access health care to allow these interventions to be that are important in Africa serotypes (1 and 5) but which implemented (United Nations General Assembly Special are not included in the current seven-valent formulation of Session on HIV/AIDS, unpublished data). Nevertheless, the vaccine, would be required to optimize the benefits of programs aimed at the prevention of the vertical transmis- the introduction of such a vaccine into Sub-Saharan Africa. sion of HIV need to be considered a priority in reducing the A challenge for other Sub-Saharan Africa countries in overall burden of LRTI among children in Sub-Saharan regard to defining the potential benefit of pneumococcal Africa. conjugate vaccination would, however, include the need to Another priority for those children who are infected with define the spectrum of serotypes that cause disease among HIV is the introduction of effective strategies aimed at children elsewhere, as geographic variation in serotypes that reducing the burden of PCP, especially considering that cause IPD have been described even within continents P. jiroveci has been found to be the etiological agent in as (Hausdorff, Siber, and Paradiso 2001). Hence, the overall many as 15 to 45 percent of African HIV-infected children proportion of IPD and possibly pneumonia that may be with LRTI. Although no randomized trial has evaluated the prevented through vaccination may differ between efficacy of trimethoprim-sulfamethoxazole (TMP-SMX) in geographic regions, depending on the prevalent serotypes. preventing PCP among children, this strategy has had pro- Despite these positive findings, and the theoretical bene- found benefits when it has been implemented in developed fits of the pneumococcal conjugate vaccine, the greatest countries as well as in such developing countries as Thailand impending limiting factor to the introduction of the vaccine (Chokephaibulkit et al. 2000; Simonds et al. 1995). Although into Sub-Saharan Africa is the cost of the vaccine. Currently recommended by the WHO and UNAIDS, data from South the vaccine is priced at $48 for the public sector in the Africa suggest that there may be structural problems with United States. Considering the costs of the vaccine and that implementing an effective TMP-SMX prophylaxis even in there are no other competing pneumococcal conjugate vac- areas with relatively good resources, such as South Africa cines that are likely to be licensed any time soon, unless (Madhi et al. 2002). Of further importance regarding the there is external funding support or major price tiering to widespread use of TMP-SMX is its potential to predispose 158 | Shabir A. Madhi and Keith P. Klugman to developing strains of pneumococcus that are resistant to disseminated tuberculosis but also pulmonary tuberculosis, it as well as to other classes of antibiotics (Madhi, Petersen, which may be responsible for 25 percent of LRTI cases Madhi, Wasas, et al. 2000). Such resistance would possibly requiring hospitalization. compromise the WHO management strategy in Sub- Saharan Africa, where in many areas, TMP-SMX remains CONCLUSION the mainstay of therapy for the treatment of LRTI. It is likely that the emergence of resistant strains of pneumococcus, Although the LRTI management strategies of the WHO coupled with further dissemination of resistant strains, may have the potential of reducing the burden of death from render TMP-SMX obsolete in the empirical management of LRTI among countries where under-five mortality is greater LRTI in Sub-Saharan Africa. than 100 per 1,000 live births, further advances in reducing The importance of TMP-SMX in reducing morbidity childhood morbidity and possibly mortality depend on the and mortality in African HIV-infected children has recently effective use of bacterial conjugate vaccines in those coun- been highlighted by a study in Malawi. Chintu and col- tries that require them, but where, unjustly for the children, leagues (2004) showed that TMP-SMX prophylaxis reduced they are least affordable. Furthermore, the reversal of gains mortality in African HIV-infected children, the majority of in reducing childhood mortality currently being experi- whom were symptomatic for AIDS, by 43 percent (hazards enced in Sub-Saharan Africa, largely a result of the HIV ratio 0.57; 95 percent CI, 0.43­0.77). epidemic, has been associated with almost a doubling of the burden of LRTI in those countries that are heavily burdened by the epidemic. An effective approach to reducing LRTI- OTHER POTENTIAL WAYS TO PREVENT LRTI associated childhood mortality in Sub-Saharan Africa requires a concerted effort to prevent the vertical transmis- Despite all the promise that the bacterial conjugate vaccines sion of HIV infection, timely and effective rollout of bacte- hold in preventing LRTI, it is sobering that the total burden rial conjugate vaccines, as well as addressing predominantly of LRTI reduced by these vaccines is relatively small, albeit poverty-linked, predisposing factors that heighten the risk important, given the overall magnitude of the burden of of children to develop severe and often fatal LRTI. LRTI in Sub-Saharan Africa. In The Gambia, the Hib con- jugate vaccine prevented only 1.3 cases of hospitalization related to LRTI for every 35.1 cases (about 4 percent) that REFERENCES occurred per 1,000 children enrolled in the study Adegbola, R. A., A. G. Falade, B. E. Sam, M. Aidoo, I. Baldeh, D. Hazlett, (Mulholland et al. 1997). The nonavalent pneumococcal H. Whittle, B. M. Greenwood, and E. K. Mulholland. 1994. "The conjugate vaccine in The Gambia reduced the overall rate Etiology of Pneumonia in Malnourished and Well-Nourished of pneumonia by 18 cases per 1,000 child-years, with an Gambian Children." Pediatric Infectious Disease Journal 13: 975­82. incidence rate of 249 per 1,000 child-years among the Afari, E. A. 1991. "Acute Respiratory Infections in Children under Five in Two Rural Communities in Southern Ghana." Japanese Journal of placebo group (Cutts et al. 2005). The case reduction among Tropical Medicine and Hygiene 19: 275­80. South African HIV-uninfected children, including only Anderson, H. R. 1978. "Respiratory Abnormalities in Papua New Guinea hospitalized children, was 2.7 per 1,000 child-years with Children: The Effects of Locality and Domestic Wood Smoke Pollution." International Journal of Epidemiology 7: 63­72. the incidence in the placebo group being 15.7 cases of LRTI Asamoah-Odel, E., G. Asiimwe-Okiror, J. M. Garcia Calleja, and J. T. per 1,000 child-years (Madhi et al. 2005). These data Boerma. 2003. "Recent Trends in HIV Prevalence among Pregnant indicate that although the most severe causes of pneumonia Women in Sub-Saharan Africa." AIDScience 3 (19). http:// aidscience.org/Articles/AIDScience037.asp. may have been prevented with modest to good success, there Bale, J. R. 1990. "Creation of a Research Program to Determine the remains a large burden of LRTI that even the expensive new Etiology and Epidemiology of Acute Respiratory Tract Infection generation of vaccines is not able to prevent. Efforts aimed among Children in Developing Countries." Reviews of Infectious Diseases 12 (Suppl. 8): S861­66. at developing vaccines against respiratory syncytial virus Berman, S., and K. McIntosh. 1985. "Selective Primary Health Care: holds the promise that greater inroads may be made in Strategies for Control of Disease in the Developing World. Part 21. reducing the overall LRTI morbidity even in Sub-Saharan Acute Respiratory Infections." Reviews of Infectious Diseases 7: 674­91. Africa countries. Furthermore, there is a need to continue Black, R. E., K. H. Brown, S. Becker, and M. Yunus. 1982. "Longitudinal Studies of Infectious Diseases and Physical Growth of Children in support for developing more effective vaccines against Rural Bangladesh. I. Patterns of Morbidity." American Journal of M. tuberculosis that are effective in preventing not only Epidemiology 115: 305­14. Acute Respiratory Infections | 159 Black, S., H. Shinefield, B. Fireman, E. Lewis, P. Ray, J. R. Hansen, L. Elvin, Fantahun, M. V. 1998. "Patterns of Childhood Mortality in Three Districts et al. 2000. "Efficacy, Safety and Immunogenicity of Heptavalent of North Gondar Administrative Zone. A Community Based Study Pneumococcal Conjugate Vaccine in Children." Northern California Using the Verbal Autopsy Method." Ethiopian Medical Journal 36: Kaiser Permanente Vaccine Study Center Group. Pediatric Infectious 71­81. Disease Journal 19: 187­95. Forgie, I. M., K. P. O'Neill, N. Lloyd-Evans, M. Leinonen, H. Campbell, Black, S. B., H. R. Shinefield, S. Ling, J. Hansen, B. Fireman, D. Spring, H. C. Whittle, and B. M. Greenwood. 1991a. "Etiology of Acute Lower J. Noyes, et al. 2002. Effectiveness of Heptavalent Pneumococcal Respiratory Tract Infections in Gambian Children: I. Acute Lower Conjugate Vaccine in Children Younger than Five Years of Age for Respiratory Tract Infections in Infants Presenting at the Hospital." Prevention of Pneumonia. Pediatric Infectious Disease Journal 21: Pediatric Infectious Disease Journal 10: 33­41. 810­15. ------. 1991b. "Etiology of Acute Lower Respiratory Tract Infections in Broome, C. V., and R. F. Breiman. 1991. "Pneumococcal Vaccine--Past, Gambian Children: II. Acute Lower Respiratory Tract Infection in Present, and Future." New England Journal of Medicine 325: 1506­8. Children Ages One to Nine Years Presenting at the Hospital." Pediatric Campbell, H., P. Byass, A. C. Lamont, I. M. Forgie, K. P. O'Neill, N. Lloyd- Infectious Disease Journal 10: 42­47. Evans, and B. M. Greenwood. 1989. "Assessment of Clinical Criteria Foy, H. M., M. K. Cooney, A. J. Maletzky, and J. T. Grayston. 1973. for Identification of Severe Acute Lower Respiratory Tract Infections "Incidence and Etiology of Pneumonia, Croup and Bronchiolitis in in Children." Lancet 1: 297­99. Preschool Children Belonging to a Prepaid Medical Care Group over a Cherian, T., K. E. Mulholland, J. B. Carlin, H. Ostensen, R. Amin, M. de Four-Year Period." American Journal of Epidemiology 97: 80­92. Campo, D. Greenberg, et al. 2005. "Standardised Interpretation of Garcia de Olarte, D., H. Trujillo, A. Uribe, and N. Agudelo. 1971. "Lung Paediatric Chest Radiographs for the Diagnosis of Pneumonia in Puncture-Aspiration as a Bacteriologic Diagnostic Procedure in Acute Epidemiological Studies." Bulletin of the World Health Organization 83: Pneumonias of Infants and Children." Clinical Pediatrics (Phila) 10: 353­59. 346­50. Chintu, C., G. J. Bhat, A. S. Walker, V. Mulenga, F. Sinyinza, K. Lishimpi, George, S. A., S. S. Bai, and A. Cherian. 1996. "Blood versus Lung Aspirate L. Farrelly, et al. 2004. "Co-Trimoxazole as Prophylaxis against Culture in Pneumonia." Indian Pediatrics 33: 871­72. Opportunistic Infections in HIV-Infected Zambian Children (CHAP): Graham, S. M., E. I. Mtitimila, H. S. Kamanga, A. L. Walsh, C. A. Hart, A Double-Blind Randomised Placebo-Controlled Trial." Lancet 364: and M. E. Molyneux. 2000. "Clinical Presentation and Outcome of 1865­71. Pneumocystis carinii Pneumonia in Malawian Children." Lancet 355: Chintu, C., C. Luo, G. Bhat, H. L. DuPont, P. Mwansa-Salamu, M. Kabika, 369­73. A. Zumla, et al. 1995. "Impact of the Human Immunodeficiency Virus Greenwood, B. 1992. "Epidemiology of Acute Lower Respiratory Tract Type-1 on Common Pediatric Illnesses in Zambia." Journal of Tropical Infections, Especially Those Due to Haemophilus influenzae Type B, in Pediatrics 41: 348­53. The Gambia, West Africa." Journal of Infectious Diseases 165 (Suppl. 1): Chokephaibulkit, K., R. Chuachoowong, T. Chotpitayasunondh, S. S26­28. Chearskul, N. Vanprapar, N. Waranawat, P. Mock, N. Shaffer, and R. J. Greenwood, B. M., A. M. Greenwood, A. K. Bradley, S. Tulloch, R. Hayes, Simonds. 2000. "Evaluating a New Strategy for Prophylaxis to Prevent and F. S. Oldfield. 1987. "Deaths in Infancy and Early Childhood in a Pneumocystis carinii Pneumonia in HIV-Exposed Infants in Thailand." Well-Vaccinated, Rural, West African Population." Annals of Tropical Bangkok Collaborative Perinatal HIV Transmission Study Group. Paediatrics 7: 91­99. AIDS 14: 1563­69. Hausdorff, W. P., G. Siber, and P. R. Paradiso. 2001. "Geographical Cutts, F. T., S. M. Zaman, G. Enwere, S. Jaffar, O. S. Levine, J. B. Okoko, Differences in Invasive Pneumococcal Disease Rates and Serotype C. Oluwalana, et al. 2005. "Efficacy of Nine-Valent Pneumococcal Frequency in Young Children." Lancet 357: 950­52. Conjugate Vaccine Against Pneumonia and Invasive Pneumococcal Disease in The Gambia: Randomised, Double-Blind, Placebo- Heiskanen-Kosma, T., M. Korppi, C. Jokinen, S. Kurki, L. Heiskanen, Controlled Trial." Lancet 365: 1139­36. H. Juvonen, S. Kallinen, et al. 1998. "Etiology of Childhood Pneumonia: Serologic Results of a Prospective, Population-Based Dagan, R., D. Fraser, Z. Greif, N. Keller, M. Kaufstein, G. Shazberg, and Study." Pediatric Infectious Disease Journal 17: 986­91. M. Schlesinger. 1998. "A Nationwide Prospective Surveillance Study in Israel to Document Pediatric Invasive Infections, with an Emphasis Hughes, W. T., R. A. Price, F. Sisko, W. S. Havron, A. G. Kafatos, on Haemophilus influenzae Type B Infections." Israeli Pediatric M. Schonland, and P. M. Smythe. 1974. "Protein-Calorie Malnutrition. Bacteremia and Meningitis Group. Pediatric Infectious Disease Journal A Host Determinant for Pneumocystis carinii Infection." American 17: S198­203. Journal of Diseases of Children 128: 44­52. Dankner, W. M., J. C. Lindsey, and M. J. Levin. 2001. "Correlates of Ikeogu, M. O. 1998. "Acute Pneumonia in Zimbabwe: Bacterial Isolates by Opportunistic Infections in Children Infected with the Human Lung Aspiration." Archives of Disease in Childhood 63: 1266­67. Immunodeficiency Virus Managed before Highly Active Antiretroviral Klugman, K. P., S. A. Madhi, R. E. Huebner, R. Kohberger, N. Mbelle, Therapy." Pediatric Infectious Disease Journal 20: 40­48. N. Pierce, and Vaccine Trialist Group. 2003. "Prevention of de Francisco, A., J. Morris, A. J. Hall, J. R. Armstrong Schellenberg, and Pneumonia and Invasive Pneumococcal Disease by a 9- B. M. Greenwood. 1993. "Risk Factors for Mortality from Acute Lower Valent Pneumococcal Conjugate Vaccine: Efficacy in Both HIV- Respiratory Tract Infections in Young Gambian Children." Inter- Infected and Uninfected Children." New England Journal of Medicine national Journal of Epidemiology 22: 1174­82. 349: 1341­48. Demers, A.-M., P. Morency, F. Mberyo-Yaah, S. Jaffar, C. Blais, P. Somse, Langston, C., E. R. Cooper, J. Goldfarb, K. A. Easley, S. Husak, S. Sunkle, G. Bobossi, and J. Pepin. 2000. "Risk Factors for Mortality among T. J. Starc, et al. 2001. "Human Immunodeficiency Virus-Related Children Hospitalized Because of Acute Respiratory Infections in Mortality in Infants and Children: Data from the Pediatric Pulmonary Bangui, Central African Republic." Pediatric Infectious Disease Journal and Cardiovascular Complications of Vertically Transmitted HIV 19: 424­32. (P(2)C(2)) Study." Pediatrics 107: 328­38. Enarson, P., and A. Pio. 2003. "Improving Pediatric Lung Health in Leowski, J. 1986. "Mortality from Acute Respiratory Infections in Children Malawi: Lessons Learned and Next Steps." Paper presented at the under 5 Years of Age: Global Estimates." World Health Statistics American Thoracic Society (ATS) Conference, Seattle, May 16­21. Quarterly 39: 138­44. 160 | Shabir A. Madhi and Keith P. Klugman Levine, O. S., R. Lagos, A. Munoz, J. Villaroel, A. M. Alvarez, P. Abrego, and in a Rural Community in The Gambia, West Africa." Pediatric M. M. Levine. 1999. "Defining the Burden of Pneumonia in Children Infectious Disease Journal 13: 122­28. Preventable by Vaccination against Haemophilus influenzae Type B." Oyejide, C. O., and K. Osinusi. 1990. "Acute Respiratory Tract Infections Pediatric Infectious Disease Journal 18: 1060­64. in Children in Idikan Community, Ibadan, Nigeria: Severity, Risk Madhi, S. A., C. Cutland, K. Ismail, C. O'Reilly, A. Mancha, and K. P. Factors and Frequency of Occurrence." Reviews of Infectious Diseases 12 Klugman. 2002. "Ineffectiveness of Trimethoprim-Sulfamethoxazole (Suppl. 8): S1042­46. Prophylaxis and Importance of Concurrent Bacterial and Respiratory Peltola, H. 2000."Worldwide Haemophilus influenzae Type B Disease at the Viral Co-Infections among Human Immunodeficiency Virus Type-1 Beginning of the 21st Century: Global Analysis of the Disease Burden Infected African Children Hospitalized with Pneumocystis carinii 25 Years after the Use of the Polysaccharide Vaccine and a Decade Pneumonia." Clinical Infectious Diseases 35: 1120­26. after the Advent of Conjugates." Clinical Microbiology Reviews 13: Madhi, S. A., C. Cutland, L. Kuwanda, and K. P. Klugman. 2005. "The 302­17. Impact of a 9-Valent Pneumococcal Conjugate Vaccine on the Public Pifer, L. L., W. T. Hughes, S. Stagno, and D. Woods. 1978. "Pneumocystis Health Burden of Pneumonia in HIV Infected and Uninfected carinii Infection: Evidence for High Prevalence in Normal and Children." Clinical Infectious Diseases 40: 1511­18. Immunosuppressed Children." Pediatrics 61: 35­41. Madhi, S. A., K. P. Klugman, and the Pneumococcal Vaccine Trialist Group. 2004. "A Role for Streptococcus pneumoniae in Virus-Associated Pillay, K., M. Colvin, R. Williams, and H. M. Coovadia. 2001. "Impact of Pneumonia." Nature Medicine 10: 811­13. HIV-1 Infection in South Africa." Archives of Disease in Childhood 85: 50­51. Madhi, S. A., H. Ludewick, Y. Abed, K. P. Klugman, and G. Boivin. 2003. "Human Metapneumovirus Associated Lower Respiratory Tract Pio, A., and P. Enarson. 2003. "Improving Child Lung Health: The Malawi Infections among Hospitalized HIV-1 Infected and HIV-1 Uninfected Child Lung Health Project Sponsored by the Bill and Melinda Gates African Infants." Clinical Infectious Diseases 37: 1705­10. Foundation." Paper presented at the American Thoracic Society (ATS) Conference, Seattle, May 16­21. Madhi, S. A., K. Petersen, A. Madhi, M. Khoolsal, and K. P. Klugman. 2000. "Burden of Disease and Drug Resistance of Bacteria in HIV Infected Prakash, J., D. K. Agraval, K. N. Agarval, and A. K. Kulati. 1996. "Etiologic Children with Severe Lower Respiratory Tract Infections." Clinical Diagnosis of Pneumonia in Under-Five Children." Indian Pediatrics 33: Infectious Diseases 31: 170­76. 329­31. Madhi, S. A., K. Petersen, A. Madhi, A. Wasas, and K. P. Klugman. 2000. Prinsloo, J. G., and A. Cicoria. 1974. "Is Lung Puncture Aspiration a "Impact of Human Immunodeficiency Virus Type-1 on the Disease Harmless Procedure?" South African Medical Journal 48: 597­98. Spectrum of Streptococcus pneumoniae in South African Children." Rapkin, R. H. 1975. "Bacteriologic and Clinical Findings in Acute Pediatric Infectious Disease Journal 19: 1141­47. Pneumonia of Childhood." Clinical Pediatrics (Phila) 14: 130­33. Madhi, S. A., B. Schoub, K. Simmank, N. Blackburn, and K. P. Klugman. Riley, I., E. Carrad, H. Gratten, M. Gratten, K. Lovuru, P. Phillips, D. Pratt, 2000. "Increased Burden of Respiratory Virus Associated Severe Lower et al. 1983. "The Status of Research on Acute Respiratory Infections in Respiratory Tract Infections in HIV-1 Infected Children." Journal of Children in Papua New Guinea." Pediatric Research 17: 1041­43. Pediatrics 137: 78­84. Riley, I. D., D. Lehmann, M. P. Alpers, T. F. Marshall, H. Gratten, and Mimica, I., E. Donoso, J. E. Howard, and G. W. Ledermann. 1971. "Lung D. Smith. 1986. "Pneumococcal Vaccine Prevents Death from Acute Puncture in the Etiological Diagnosis of Pneumonia. A Study of 543 Lower-Respiratory-Tract Infections in Papua New Guinean Children." Infants and Children." American Journal of Diseases of Children 122: Lancet 2: 877­81. 278­82. Rudan, I., L. Tomaskovic, C. Boschi-Pinto, and H. Campbell. 2004. "Global Mofenson, L. M., R. Yogev, J. Korelitz, J. Bethel, K. Krasinski, J. Moye Jr., Estimate of the Incidence of Clinical Pneumonia among Children R. Nugent, and J. G. Rigau-Perez. 1998. "Characteristics of Acute under Five Years of Age." Bulletin of the World Health Organization 82: Pneumonia in Human Immunodeficiency Virus-Infected Children 895­903. and Association with Long Term Mortality Risk." National Institute of Child Health and Human Development Intravenous Immunoglobulin ------. Forthcoming. "Incidence of Acute Lower Respiratory Infections Clinical Trial Study Group. Pediatric Infectious Disease Journal 17: (ALRI) in Children under Five Years: II Estimates for WHO Regions." 872­80. Bulletin of the World Health Organization. Mtango, F. D., and D. Neuvians. 1986. "Acute Respiratory Infections in Sazawal, S., and R. E. Black. 2003. "Pneumonia Case Management Trials Children under Five Years." Control project in Bagamoyo District, Group Effect of Pneumonia Case Management on Mortality in Tanzania. Transactions of the Royal Society of Tropical Medicine and Neonates, Infants, and Preschool Children: A Meta-Analysis of Hygiene 80: 851­58. Community-Based Trials." Lancet Infectious Diseases 3: 547­56. Mulholland, K., S. Hilton, R. Adegbola, S. Usen, A. Oparaugo, C. Selwyn, B. J. 1990."The Epidemiology of Acute Respiratory Tract Infection Omosigho, M. Weber, et al. 1997. "Randomised Trial of Haemophilus in Young Children: Comparison of Findings from Several Developing influenzae Type-B Tetanus Protein Conjugate Vaccine." Lancet 349: Countries." Coordinated Data Group of BOSTID Researchers. Reviews 1191­97. of Infectious Diseases 12 (Suppl. 8): S870­88. Murphy, T. F., F. W. Henderson, W. A. Clyde Jr., A. M. Collier, and Silverman, M., D. Stratton, A. Diallo, and L. J. Egler. 1977. "Diagnosis F. W. Denny. 1981. "Pneumonia: An Eleven-Year Study in a Pediatric of Acute Bacterial Pneumonia in Nigerian Children. Value of Needle Practice." American Journal of Epidemiology 113: 12­21. Aspiration of Lung of Countercurrent Immunoelectrophoresis." Nathoo, K. J., S. Chigonde, M. Nhembe, M. H. Ali, and P. R. Mason. 1996. Archives of Disease in Childhood 52: 925­31. "Community-Acquired Bacteremia in Human Immunodeficiency Simonds, R. J., M. L. Lindegren, P. Thomas, D. Hanson, B. Caldwell, G. Virus-Infected Children in Harare, Zimbabwe." Pediatric Infectious Scott, and M. Rogers. 1995. "Prophylaxis against Pneumocystis carinii Disease Journal 15: 1092­97. Pneumonia among Children with Perinatally Acquired Human O'Dempsey, T. J., T. F. McArdle, N. Lloyd-Evans, I. Baldeh, B. E. Laurence, Immunodeficiency Virus Infection in the United States." Pneumocystis O. Secka, and B. M. Greenwood. 1994. "Importance of Enteric Bacteria carinii Pneumonia Prophylaxis Evaluation Working Group. New as a Cause of Pneumonia, Meningitis and Septicemia among Children England Journal of Medicine 332: 786­90. Acute Respiratory Infections | 161 Sofoluwe, G. O. 1968. "Smoke Pollution in Dwellings of Infants with Walker, N., B. Schwartlander, and J. Bryce. 2002. "Meeting International Bronchopneumonia." Archives of Environmental Health 16: 670­72. Goals in Child Survival and HIV/AIDS." Lancet 360: 284­89. Takala, A. K., J. Eskola, H. Peltola, and P. H. Makela. 1989. "Epidemiology Wall, R. A., P. T. Corrah, D. C. Mabey, and B. M. Greenwood. 1986. "The of Invasive Haemophilus influenzae Type B Disease among Children Etiology of Lobar Pneumonia in The Gambia." Bulletin of the World in Finland before Vaccination with Haemophilus influenzae Type B Health Organization 64: 553­58. Conjugate Vaccine." Pediatric Infectious Disease Journal 8: 297­302. WHO (World Health Organization). 1990. Management of the Young Child UNICEF (United Nations Children's Fund). 1996. The State of the World's with Acute Lower Respiratory Infection. WHO Programme for the Children 1996. Statistical Tables. Geneva: UNICEF. Control of Acute Respiratory Infections. Geneva: WHO. van den Hoogen, B. G., J. C. de Jong, J. Groen, T. Kuiken, R. de Groot, Williams, B. G. 2002. "Estimates of World-Wide Distribution of Child R. A. Fouchier, and A. D. Osterhaus. 2001. "A Newly Discovered Deaths from Acute Respiratory Tract Infections." Lancet Infectious Human Pneumovirus Isolated from Young Children with Respiratory Diseases 2: 25­32. Tract Disease." Nature Medicine 7: 719­24. Wubbel, L., L. Muniz, A. Ahmed, M. Trujillo, C. Carubelli, C. McCoig, Von Schirnding, Y. E. R., D. Yach, and M. Klein. 1991 "Acute Respiratory T. Abramo, M. Leinonen, and G. H. McCracken Jr. 1999. "Etiology and Tract Infections as an Important Cause of Childhood Deaths in South Treatment of Community-Acquired Pneumonia in Ambulatory Africa." South African Medical Journal 80: 79­82. Children." Pediatric Infectious Disease Journal 18: 98­104. Vuori-Holopainen, E., and H. Peltola. 2001. "Reappraisal of Lung Tap: Zangwill, K. M., C. M. Vadheim, A. M. Vannier, L. S. Hemenway, D. P. Review of an Old Method for Better Etiologic Diagnosis of Childhood Greenberg, and J. I. Ward. 1996. "Epidemiology of Invasive Pneumonia." Clinical Infectious Diseases 32: 715­26. Pneumococcal Disease in Southern California: Implications for the Vuori-Holopainen, E., E. Salo, H. Saxen, K. Hedman, T. Hyypia, Design and Conduct of a Pneumococcal Conjugate Vaccine Efficacy R. Lahdenpera, M. Leinonen, E. Tarkka, M. Vaara, and H. Peltola. 2002. Trial." Journal of Infectious Diseases 174: 752­59. "Etiological Diagnosis of Childhood Pneumonia by Use of Zar, H. J., D. Hanslo, E. Tannenbaum, M. Klein, A. Argent, B. Eley, Transthoracic Needle Aspiration and Modern Microbiological J. Burgess, K. Magnus, E. D. Bateman, and G. Hussey. 2001. "Aetiology Methods." Clinical Infectious Diseases 34: 583­90. and Outcome of Pneumonia in Human Immunodeficiency Virus- Wafula, E. M., F. E. Onyango, W. M. Mirza, W. M. Macharia, I. Wamola, Infected Children Hospitalized in South Africa." Acta Paediatrica 90: J. O. Ndinya-Achola, R. Agwanda, R. N. Waigwa, and J. Musia. 1990. 119­25. "Epidemiology of Acute Respiratory Tract Infections among Young Zwi, K. J., J. M. Pettifor, and N. Soderlund. 1999. "Paediatric Hospital Children in Kenya." Reviews of Infectious Diseases 12 (Suppl. 8): Admissions at a South African Urban Regional Hospital: The Impact S1035­38. of HIV, 1992­1997." Annals of Tropical Paediatrics 19: 135­42. 162 | Shabir A. Madhi and Keith P. Klugman Chapter 12 Vaccine-Preventable Diseases Mark A. Miller and John T. Sentz Vaccines have been frequently cited as one of the most equi- which grew out of the smallpox eradication initiative, have table low-cost, high-impact public health measures, saving developed in many countries through the administrative, millions of lives annually when programs are implemented technical, and financial support of the United Nations on the national level. Over the last 40 years, the use of Children's Fund (UNICEF), the World Health Organization smallpox, measles, diphtheria, tetanus, pertussis, and (WHO), and many bilateral or multilateral partner agencies poliomyelitis vaccines have eradicated smallpox and elimi- (figure 12.1) (WHO and UNICEF 1996). In its 1993 World nated disease in those populations that have achieved and Development Report, the World Bank classified vaccination sustained programs with high implementation rates. as one of the most cost-effective public health interventions Although there are numerous licensed vaccines that could (World Bank 1993). In addition, vaccination programs have potentially benefit the African population, only those been cited as providing one of the most equitable of public routinely used and potential vaccines with broad applica- health programs, providing protection to the entire popula- tion on the horizon are covered in this chapter. tion when successfully implemented. The WHO created the Expanded Program on Immunization (EPI) in 1974 as a means to continue the HISTORY OF VACCINATION IN AFRICA great success that had been achieved earlier with the eradi- cation of smallpox. At that time less than 5 percent of the The eradication of smallpox was an outstanding display of world's children in the developing world were receiving concerted global action in a war against microbial invaders. immunizations. The six diseases chosen to be tackled under The progress in expanding poliomyelitis and measles vacci- this new initiative were tuberculosis, diphtheria, tetanus, nation efforts and their elimination from many regions pertussis, polio, and measles. It was not until 1988 that the further demonstrates that vaccines are among the most WHO recommended that yellow fever vaccine be added to powerful public health tools. National vaccination programs, the national immunization programs of those countries 163 Figure 12.1 Percentage of Target Population in Africa Table 12.1 WHO-Estimated Deaths and DALYs from Vaccinated, by Vaccine Type, 1980­2002 Vaccine-Preventable Diseases, 2002 (thousands) 80 70 Deaths DALYs 60 Total Africaa Total Africaa 50 EPI Vaccines population 40 Tuberculosis 1,566 348 34,736 9,266 target 30 Childhood-cluster diseases 1,124 527 41,480 18,995 of % 20 Diphtheria 5 2 185 48 10 Tetanus 214 84 7,074 2,775 0 Pertussis 294 131 12,595 5,243 1980 1990 1998 1999 2000 2001 2002 2003 year Poliomyelitis 1 0 151 15 Measles 611 311 21,475 10,915 BCG DTP3 HBV Measles Polio TT YF EPI Plus Vaccines Hepatitis B Source: WHO 2003. Note: BCG bacillus Calmette-Guérin; DTP3 three doses of diphtheria, tetanus, Acute hepatitis B 103 20 2,170 582 pertussis vaccine; HBV hepatitis B virus; TT tetanus toxoid; and YF yellow fever. Liver cancer 618 45 7,135 90 Cirrhosis of the liver 786 54 13,977 957 with endemic disease (WHO and UNICEF 1996). Later, in Meningitis 173 20 6,192 891 1992, the World Health Assembly recommended hepatitis B Lower respiratory infections 3,884 1,104 91,374 34,911 vaccination for all infants. Most recently the WHO has Otitis media 4 2 1,435 263 recommended that the Haemophilus influenzae type B (Hib) Source: WHO 2004. conjugate vaccines be implemented into national immu- a. WHO designated African countries. nization programs unless epidemiological evidence exists of low disease burden, lack of benefit, or overwhelming obsta- a national level can help to account for causes of competing cles to implementation (WHO 2006). mortality, allowing the assessment of health outcomes as deaths, years of life lost, or other measures (for example, DALYs). Estimates of disease burden have evolved over the THE VACCINE-PREVENTABLE DISEASE BURDEN years from simplistic models incorporating reported rates of disease and a reporting efficiency profile to the more com- The WHO-estimated disease burdens from vaccine- plex Susceptible-Infected-Removed (SIR)­type models that preventable diseases are shown for 2002 by incidence of can additionally account for local population-based cover- death and disability-adjusted life years (DALYS) in table 12.1. age data and other deterministic data, such as socioeconomic As the WHO does not necessarily classify disease based on factors and characteristics of the heterogeneity of popula- what may be prevented by a specific vaccine, the table tions throughout the region. Models are based on numerous includes those syndromes that may be partially mitigated assumptions, and their degree of accuracy is only as good as by vaccines (for example, acute lower respiratory tract infec- the data that support the many assumptions. Although tion, meningitis, or conditions associated with hepatitis B disease burden is frequently indicated as point estimates, infections). it is more appropriate to indicate the burden by a range of As most vaccine-preventable diseases are underreported values to reflect uncertainty, sometimes by an order of in many countries, estimates of disease burden are made by magnitude or more for certain diseases. a variety of methodologies that account for the susceptible fraction of the population, as calculated from natural Polio immunity from presumed historical infections, historical immunization coverage rates, and vaccine effectiveness. Poliovirus is most often transmitted fecal-orally among Disease burden estimates also integrate rates of infectivity, persons living in unsanitary and crowded conditions. Acute specific sequelae, and local case fatality. Life expectancies on infections are caused by any one of three serotypes of 164 | Mark A. Miller and John T. Sentz poliovirus that initially replicate in the gastrointestinal tract. per 3.3 million doses administered (Sutter and Kew 2004). Exposure to poliovirus predominantly results in asympto- Viral shedding can be chronic and occur for years. Recent matic infections. It has been estimated that 24 percent of polio outbreaks have been blamed on continuous transmis- infections result in minor illness characterized by a few days sion of polioviruses derived from vaccine strains, most of varying symptoms, including fever, malaise, drowsiness, notably in areas with low immunization coverage rates headache, nausea, vomiting, constipation, and sore throat (Kew et al. 2004). An assumption of the WHO eradication (Gelfand et al. 1957). In fewer cases (4 percent), infection program is that countries would be able to stop vaccination leads to nonparalytic polio or aseptic meningitis, which after wild poliovirus ceases transmission; however, the pos- manifests as fever, vomiting, malaise, and sore throat; sible threat of a resurgence because of the continued likely meningeal irritation occurs one to two days later, character- circulation of oral poliovirus vaccine would seriously chal- ized by soreness and stiffness of the neck, back, limbs, and lenge that strategy (Miller, Barrett, and Henderson 2006). severe headache (Horstmann 1955). These symptoms can Inactivated poliovirus vaccine (IPV) is another vaccine, last up to 10 days, but recovery is usually rapid and com- composed of three types of inactivated polioviruses. When plete. Paralytic polio, which affects less than 1 percent of IPV is administered to infants as the primary vaccine, the those infected, is the most serious manifestation of the dis- first two or three doses should be administered in the first ease (Sabin 1951). This form of the disease presents initially six months of life, followed by a booster during the second as a minor fever with rapid progression to paralysis within a year. It is also recommended that, when feasible, a second matter of days. Paralysis may affect the major muscles booster be given to children before they enter school. IPV involved in respiration and therefore cause death if there is efficacy rates have been estimated between 80 and 90 per- no appropriate rapid intervention. cent against paralytic polio (Plotkin and Vidor 2004). An Global efforts toward polio eradication have included increasing number of countries are transitioning from vaccination campaigns and active surveillance. The annual OPV to IPV because of the risk intolerance associated with incidence of paralytic polio was reduced from an estimated vaccine-derived paralysis. Financial and operational barriers 350,000 in 1988 to about 1,000 from 2001 to 2004 (WHO currently prevent many developing countries from doing 2004). Africa and South Asia are the last regions in the world the same, although costs and the ease of IPV use through where poliomyelitis is still endemic. False accusations of combination vaccines may allow their greater use in the tainted vaccines by local leaders have led to a local resur- future. gence of poliomyelitis cases and consequent spread to other parts of Africa (Heymann and Aylward 2004). Measles Two different poliomyelitis vaccines are currently used to protect against disease. Oral poliovirus vaccine (OPV) is a Measles is an acute, highly infectious viral disease that is live attenuated vaccine containing all three serotypes of the transmitted from person to person through large respira- poliovirus and induces protection as high as 95 percent in tory droplets. In the absence of vaccination, measles is esti- individuals who receive three doses. Additional booster mated to infect virtually the entire population with the doses are necessary to achieve nearly 100 percent protection. exception of isolated communities (Black 1976). Most chil- Currently the WHO recommends that OPV be given at dren born to immune mothers are protected from the virus birth, 6, 10, and 14 weeks in polio endemic or recently for the first six months of their lives from acquired maternal endemic countries (Sutter and Kew 2004). In many devel- antibodies. More than 90 percent of infections are associ- oped countries this vaccine is now delivered in a series of ated with clinical disease, which includes a range of symp- national or subnational campaigns several times a year. The toms, including fever, rhinorrhea, cough, and conjunctivitis. majority of developing countries throughout the world rely A rash commonly appears within three to four days after on OPV for vaccinating their population, a combined birth onset. Complications associated with measles include pneu- cohort of 127 million people (Sutter and Kew 2004). monia, diarrhea, encephalitis, and blindness. The case-fatality Because OPV is a live viral vaccine that can revert to a rates in recent years have been estimated at 0.1 to 3 percent transmissible pathogenic virus, recipients or those in con- in many developing nations (Strebel, Papania, and Halsey tact with them can be at increased risk of vaccine-associated 2004). paralytic poliomyelitis (VAPP) due to shedding of live virus The number of deaths due to measles has been a subject into the environment. VAPP occurs in less than one person of considerable controversy for the past several years, mostly Vaccine-Preventable Diseases | 165 because of the inability to specify accurately the cause of Figure 12.2 Immunization Coverage with Measles-Containing death in children afflicted with measles and other, similar Vaccines, 2003 conditions. Many models have been constructed that (percent) demonstrate a substantial reduction of measles deaths from 30° 20° 10° 0° 10° 20° 30° 40° 50° IBRD 34591 MARCH 2006 the long-stated global number of 1 million deaths per year Mediterranean Sea reported by the WHO prior to 1997. Large-scale urban and 30° 30° nationwide vaccination campaigns over the last few years Red have reduced measles mortality to 250,000­500,000 deaths 20° Sea 20° per year, most of which still occur in Sub-Saharan Africa. Measles can coexist with several comorbid conditions that 10° 10° can cause death, so attribution to any one cause is somewhat arbitrary, rendering specific accounting difficult. It is likely, 0° ATLANTIC 0° however, that recent implementation of mass vaccination OCEAN INDIAN OCEAN campaigns throughout many regions, and in areas previ- 10° ously associated with high mortality, has dramatically 10° reduced deaths due to measles. However, given the high transmission rate of measles virus, this reduction in mortal- 20° 20° < 50 ity could only be maintained with continuous high cover- 50 ­ 79 80 ­ 90 age. Measles mortality can be estimated from the suscepti- > 90 30° ­ 30° no data bility profile of the population based on historical immu- international boundaries This map was produced by the Map Design Unit of The World Bank. nization coverage rates and natural immunity (Miller 2000). The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. Transmission can be blocked if population-based immunity 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° exceeds approximately 93 percent, limiting cases only to Source: WHO and UNICEF 2003. importations. Control in many urban parts of Africa may be difficult, given that transmission is higher in densely popu- low-cost health interventions, low coverage rates of the vac- lated environments with low levels of hygiene (Miller 2000). cine in the 1990s is indicative of the poor state of public The measles vaccine is a live attenuated vaccine that can health infrastructures in various Sub-Saharan Africa popu- be administered by itself, combined with rubella vaccine, or lations (figure 12.2). with mumps and rubella vaccines. Nearly 95 percent of chil- dren vaccinated with at least one combination of the vaccine Diphtheria-Tetanus-Pertussis develop immunity. Most individuals who fail to develop immunity with one dose will do so upon receiving a second Diphtheria, tetanus, and pertussis are frequently cited dose. Vaccination is believed to induce lifelong immunity. together given that these three diseases are frequently con- The first dose should be given on or after the first birthday. trolled with a single vaccine. Each of these diseases will be A second dose is given to children when they are about four covered in turn. to six years old but can be given as soon as one month after the first dose. Adverse reactions following immunization, Diphtheria. Diphtheria is caused by toxin-producing such as fever, rash, or lowered platelet counts, have been strains of the bacterium Corynebacterium diphtheriae, observed in 5 to 15 percent of vaccine recipients. Many which can be transmitted from person to person via respi- developing countries are now delivering measles vaccines in ratory droplets. The bacterium often affects the tonsils, campaign style, similar to that used to deliver poliovirus pharynx, nasal mucosa, inner ear, vagina, or skin. Respi- vaccination. ratory diphtheria presents as a sore throat often accompa- Measles vaccination rates are sensitive indicators of func- nying a mild fever. Death can result from severe cases in tional public health systems. One to two doses of a US$0.14 which swelling from pharyngeal and tracheal exudates vaccine could prevent a disease that practically affects obstruct the airway. Myocarditis and neuritis are two other 100 percent of the population with an approximate case- complications associated with respiratory diphtheria. fatality rate of 3 to 6 percent in Sub-Saharan Africa. Given Cutaneous diphtheria presents as skin lesions and causes far that measles vaccine is one of the most cost-effective and fewer complications and deaths among those infected. 166 | Mark A. Miller and John T. Sentz Estimates of the burden of diphtheria in Africa are unre- Pertussis is characterized as spasms (paroxysms) of cough- liable, given the low number of diagnosed or reported ing followed by inspiratory "whooping." The paroxysms can cases. In The Gambia, an annual incidence rate of six cases vary in length and severity but may become so severe, espe- per 1,000 persons under the age of five years was reported cially among infants, that respiration is compromised, (Heyworth and Ropp 1973). Incidence data in Africa is resulting in hypoxia. In some cases this can cause neurolog- limited to case series and hospital-based surveillance stud- ical damage. Pneumonia can also be a complication of per- ies, where underreporting is likely, given that diphtheria is tussis infection. In a study conducted in Canada, pneumo- frequently reported as nonspecific upper respiratory infec- nia occurred in 2 percent of patients younger than 30 years tions (Rodrigues 1991). old and in 5 to 9 percent of older participants (De Serres The EPI has traditionally recommended three doses of et al. 2000). Severe coughing in older persons can cause the combined diphtheria-tetanus-pertussis (DTP) vaccine serious complications, ranging from rib fractures to pneu- in the first year of life (in conjunction with polio vaccine). mo-thorax, inguinal hernia, and herniated lumbar disks Most developed countries give subsequent booster doses in (De Serres et al. 2000; Postels-Multani et al. 1995). childhood and diphtheria-tetanus boosters in adulthood. In more developed countries transmission from adults to There is growing concern that adolescents and adult popu- young infants is common. Girls tend to have higher inci- lations are becoming more susceptible to diphtheria because dence rates of the disease than boys (Dragsted et al. 2004; repeated doses of diphtheria toxoid are needed to maintain Mahieu et al. 1978; Preziosi et al. 2002). Pertussis is highly immunity in these populations. Once a child is immunized, contagious in its early stages and has a secondary attack rate the immunity wanes relatively rapidly without exposure in other household members as high as 90 percent to natural C. diphtheriae (Galazka and Robertson 1996; (Rodrigues 1991). Reported incidence rates of pertussis in Galazka, Robertson, and Oblapenko 1995; Geldermalsen Senegal prior to any immunization campaigns were esti- and Wenning 1993). However, in Africa, the need for boost- mated to be 183 per 1,000 child-years at risk under age five ers is offset by the natural immunity provided by the pres- years, with a 2.8 percent case-fatality rate (Preziosi 2002). ence of C. diphtheriae in skin ulcers as well as asymptomatic Similar studies have shown incidence rates in Africa to be carriage in the throat, which spreads the organism through- nearly 1,000 per 100,000 inhabitants (Galazka 1992). out the population. Carrier rates in Africa have been esti- Each year there are an estimated 20 million to 40 million mated to be as high as 9.3 percent in children in the general cases of pertussis and another 200,000 to 400,000 deaths population (Geldermalsen and Wenning 1993). attributed to the disease, 90 percent of which occur in The risk of diphtheria epidemics is heightened among the developing world (WHO 1999b). The WHO believes communities with an immunity gap in adults and a large that only 1 to 2 percent of cases are reported worldwide. number of susceptible children and adolescents (Galazka Pertussis diagnosis is difficult for several reasons. Paroxysms and Robertson 1996). As a means of controlling potential among adults are less severe and often misdiagnosed as diphtheria outbreaks, immunization coverage rates should other respiratory illnesses. Misdiagnosis is common in areas be increased among at-risk groups, cases should be without adequately trained personnel or technology. In a promptly detected and managed, and close contacts should study of 3,096 patients, B. pertussis was found in 496 indi- be recognized quickly to prevent secondary infections viduals; 208 (42 percent) were diagnosed by polymerase (Galazka, Robertson, and Oblapenko 1995). For these rea- chain reaction (PCR) alone, whereas 17 (3 percent) were sons effective vaccine and surveillance must be maintained diagnosed by culture alone (Dragsted et al. 2004). Rapid in Sub-Saharan Africa in order to keep the prevalence of diagnosis of pertussis using PCR techniques together with diphtheria at relatively low levels and to prevent possible serological assays can enhance diagnosis as well as surveil- epidemics. In addition, the use of booster doses should be lance of pertussis (Fry et al. 2004). Clinical diagnosis of per- considered, especially as immunity wanes in those popula- tussis by a trained physician has also proved to be a reliable tions with increasing hygiene and the consequent decrease diagnostic tool (Granstrom, Wretlind, and Granstrom 1991) to natural reexposure. and is often characterized by a cough that lasts at least 14 days (Patriarca et al. 1988). These are important implica- Pertussis. Pertussis, or whooping cough, is a highly conta- tions when considering diagnosis of pertussis in remote gious disease caused by the bacterium Bordetella pertussis, areas with limited laboratory resources and few trained which is transmitted through respiratory excretions. health professionals. Vaccine-Preventable Diseases | 167 The incidence rate of pertussis has declined drastically mortality rates greater than 5 per 1,000 live births are in over the past half-century primarily because of the admin- Africa (WHO 1999c). istration of the inactivated whole-cell pertussis vaccines. The WHO estimates that approximately 3 percent of Due to neurological reactions associated with the whole- NNT cases are reported each year; this figure includes those cell vaccines, new acellular vaccines have been developed. reported from countries considered to have well-developed Either of these vaccines is usually administered with the surveillance systems (WHO 1999b). Because death may diphtheria and tetanus toxoids (TTs). The whole-cell vac- occur within the first week of birth, it often is unreported in cine is cheaper than the acellular vaccine and is produced official mortality records and therefore is referred to as "the in many developing countries (WHO 1999b). A herd effect silent killer" (CVI 1994). Community surveys conducted in of vaccination not only protects immunized infants but various states of Nigeria between 1990 and 1993 found that decreases transmission rates to protect unvaccinated mortality rates from NNT ranged between 9 and 20 per infants (Miller and Gay 1997). The introduction of pertus- 1,000 live births (Babaniyi and Parakoyi 1991). It was esti- sis vaccines through the EPI in Senegal has resulted in mated in 1999 that nearly 46,000 cases of NNT occurred in steady and dramatic decreases in the incidence of the dis- Nigeria, representing 36 percent of all the NNT cases in the ease, especially among infants under six months old African region, although only 1,529 of the cases were actu- (Preziosi et al. 2002). ally reported. This problem of underreporting has serious implications for the control of the disease. Tetanus. Tetanus is the only EPI vaccine-preventable dis- The World Health Assembly in 1989 set the goal for elim- ease that is not communicable but acquired through ination (defined as incidence of less than 1 case per 1,000 environmental contamination. The bacterium Clostridium live births) of NNT by 1995. This date was later delayed until tetani, which can grow in dirty flesh wounds, produces a 2005 because of operational constraints from the expanded neurotoxin causing convulsions and eventual death. scope of including maternal tetanus. By 1999 only 104 of Neonatal tetanus (NNT), the most common form of the 161 developing countries in the world had eliminated tetanus in the developing world, is the result of contamina- neonatal tetanus. Sixteen of the 57 nations that have yet to tion of the umbilical stump either by the use of nonsterile reach this target are located in the African region (fig- instruments after delivery or the application of animal ure 12.3) (Idema et al. 2002). Included in this plan for elim- dung to the cut cord, a custom in many cultures, especially ination is the use of TT to protect pregnant women and the among groups in Sub-Saharan Africa (Elmore-Meegan use of supplemental immunization activities, clean birthing et al. 2001). Symptoms can appear 3 to 14 days after birth practices, and active surveillance systems for women of following a period of normal feeding. Infected infants childbearing age (WHO 2001b). will gradually lose the ability to nurse properly followed by a period of convulsions, which increase in intensity and DTP Vaccines. Inactivated diphtheria and TTs can be com- frequency. Mortality rates of infants can range from 25 to bined with whole-cell or acellular pertussis vaccines, such as 90 percent with care, 95 percent without (http://www.who. DTwP or DTaP, respectively, and administered as a single int/vaccines/en/neotetanus.shtml). injection. In most of Africa, the national EPI schedule of NNT, which causes an estimated 450,000 infant deaths, is vaccination is at 6, 10, and 14 weeks of life. Although most defined as tetanus in the first month of life. Another 40,000 developed countries administer booster doses, this does not maternal deaths are estimated to occur from tetanus routinely occur in Africa. Pertussis vaccine efficacy is 70 acquired during delivery (WHO 1999a). NNT remains a to 90 percent of fully vaccinated children (Edwards and global problem, but the greatest burden occurs in Africa Decker 2004); however, continuous protection requires (table 12.1) (Anita-Obong, Young, and Effiong 1993; booster doses. Three doses of tetanus or diphtheria toxoids Stanfield and Galazka 1984; WHO 1999c, 2001b). Nearly are associated with greater than 95 percent protection 90 percent of the global burden from NNT is from 28 coun- against disease (Cherry and Harrison 2004; Feigin, tries. Of these, 16 are located in Africa (Angola, Burkina Stechenberg, and Hertel 2004; Wharton and Vitek 2004); Faso, Cameroon, Chad, Côte d'Ivoire, Democratic Republic however, periodic boosting is necessary to ensure lifelong of Congo, Ethiopia, Ghana, Guinea-Bissau, Liberia, Mali, protection. DTP vaccine has been associated with adverse Mauritania, Mozambique, Niger, Nigeria, and Senegal). effects, such as local swelling and tenderness, fever, Additionally, 11 of the 12 countries reporting NNT febrile seizures, anaphylaxis, hyporesponsive episodes, and 168 | Mark A. Miller and John T. Sentz Figure 12.3 Maternal and Neonatal Tetanus Elimination be passed from the mother to the fetus. In addition to being Status, 2002 an effective vaccine, it is relatively cheap at about US$0.07 per dose (CVI 1994). Studies have shown that when 80 per- 30° 20° 10° 0° 10° 20° 30° 40° 50° IBRD 34592 MARCH 2006 cent of women in an area have been vaccinated against Mediterranean Sea tetanus with two doses, the level of seroprotection coupled 30° 30° with health education is adequate to eliminate tetanus Red (CVI 1994). 20° Sea 20° Some of the lowest coverage rates of TT are found in Africa. The EPI coverage rates among pregnant women with 10° 10° two doses had stagnated between 30 and 40 percent in most African nations by 1999. One of the objectives of the EPI is 0° ATLANTIC 0° to increase coverage of three doses of DTP vaccine to OCEAN INDIAN 80 percent in all districts in the region. It is estimated that OCEAN 53 percent of children under the age of one living in Sub- 10° 10° Saharan Africa have received three doses of the DTP vaccine (figure 12.4) (World Bank 2003). 20° 20° MNT eliminated MNT eliminated from over 90% of districts MNT eliminated from over 50 ­ 90% of districts Tuberculosis 30° MNT eliminated from less than 50% of districts 30° no data international boundaries Tuberculosis (TB) is a bacterial infection caused by This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° Mycobacterium tuberculosis. Transmitted through respira- Source: WHO and UNICEF 2003. tory droplets, tuberculosis is highly contagious, with studies showing a 25 to 50 percent infection rate of those in close contact with infected individuals (Smith and Starke 2004). encephalopathy (Edwards and Decker 2004). The more TB presents clinically with a wide range of symptoms, expensive acellular pertussis vaccines have greatly reduced depending upon the age of the individual. Infants and ado- these reactions. lescents are most likely to have significant clinical manifes- The TT, given to immunize pregnant women to prevent tations compared with older children. Most infants will neonatal and maternal tetanus, induces antibodies that can present with a nonproductive cough, shortness of breath, Figure 12.4 Global and Regional Immunization Coverage, Three Doses DTP, 1980­2001 100 90 80 70 60 50 percent 40 30 20 10 0 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 year global industrial countries Latin America and the Caribbean South Asia Central Europe East Asia and Pacific Middle East and North Africa Sub-Saharan Africa Source: WHO and UNICEF 2003. Vaccine-Preventable Diseases | 169 and low grade fever. Other common symptoms include Figure 12.5 Epidemics of Yellow Fever in Africa Reported to night sweats, malaise, irritability, fatigue, and weight loss. TB the WHO, 1980­2003 is often misdiagnosed as other illnesses, such as bronchitis, 30° 20° 10° 0° 10° 20° 30° 40° 50° IBRD 34527 MARCH 2006 until it progresses to more advanced stages. TB can also Mediterranean Sea present as meningitis in infants or cause a chronic infectious 30° 30° process almost anywhere throughout the body. The time between preclinical infection and the onset of disease can be Red 20° Sea 20° several weeks to many decades. Adults who are able to sup- press acute infections often carry the bacterium latently, 10° 10° which results in reactivation of the disease later in life. May July 2003 2003 Accurate diagnosis of TB is a critical component for better 0° 0° understanding the burden of disease as well as the effective- INDIAN ness of interventions. In resource-poor areas of the world in ATLANTIC OCEAN OCEAN which mycobacterial cultures and radiography are not 10° 10° accessible, microscopy is most commonly used to identify the organism and thus diagnose the disease. 20° 20° TB is blamed for nearly 2 million to 3 million deaths countries at risk annually, and it is believed that another 8 million people are countries notifying YF epidemic 1980 ­ 1999 30° 30° countries notifying epidemic since 2000 infected with the bacterium each year. Nearly one-third of no data international boundaries This map was produced by the Map Design Unit of The World Bank. the world's population is currently infected with TB (WHO The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. 2001a). In much of the world the TB incidence rates contin- 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° ue to grow, especially in Sub-Saharan Africa, despite the Source: WHO 2003. widespread use of the bacille Calmette-Guérin (BCG) vac- cine (Cantwell and Binkin 1996). In Africa, coinfections of Yellow fever is endemic to parts of South America as human immunodeficiency virus (HIV) and TB have led to well as Sub-Saharan Africa in areas bordering jungles increases in the incidence rate of TB by approximately (figure 12.5). Between 1986 and 1995, reported incidence of 20 percent (Smith and Starke 2004). yellow fever dramatically increased from previous reporting BCG is a live attenuated bacterial vaccine most commonly intervals, likely because of cessation of vaccinations. In administered intradermally at birth to prevent tuberculosis. Africa alone, 22,952 cases were reported, accounting for The effectiveness of the BCG vaccine against TB has been 89 percent of the total global cases during that period debated, with a range estimated from 0 to 80 percent (Fine (Monath 2004). A case-fatality rate of 23 percent (5,357 2001). Most proponents claim that it is effective against TB deaths) was reported throughout Africa. meningitis, but it is not commonly believed to prevent TB in The yellow fever vaccine is a live attenuated viral vaccine adults nor its transmission. The real impact of BCG may recommended for anyone nine years or older living or trav- have been confounded by many other improvements in eling to endemic regions of South America and Africa. A public health that could have contributed to the decrease in single dose of the vaccine has shown to be effective for 30 to disease burden associated with tuberculosis (Smith and 35 years and most likely for the duration of the vaccine Starke 2004). recipient's life (Monath 2004). Adverse effects from the vac- cine have ranged from mild symptoms of headache, malaise, or low-grade fever to rare, more serious complications. Yellow Fever Yellow fever is an acute viral infection transmitted by mos- Hepatitis B quito, primarily the Aedes aegypti. Symptoms of the infec- tion can vary, making diagnosis difficult and leading to the Many viral agents cause hepatitis; vaccine can counteract underestimation of morbidity (Monath 2004). After an two of them--hepatitis A virus (HAV) and hepatitis B virus incubation period of three to six days, onset of symptoms-- (HBV). Although a licensed vaccine against HAV is used in rigors, headache, nausea, joint pain, jaundice, and myalgia-- developed countries, it is currently not considered to be is rapid. In approximately 15 to 20 percent of cases, severe cost-effective for Africa. HBV, transmitted through blood- disease causes multiple organ failure. borne infections, sexually, or from mother to infant, can 170 | Mark A. Miller and John T. Sentz potentially cause more severe illness, including fulminant Figure 12.6 African Nations Using Hepatitis B Vaccine in a hepatitis, cirrhosis, and liver cancer. Persons infected as National Infant Immunization Campaign, 2002 infants have a 70 to 90 percent chance of becoming a 30° 20° 10° 0° 10° 20° 30° 40° 50° IBRD 34593 MARCH 2006 chronic carrier of the virus and can then either infect other Mediterranean Sea persons or develop severe sequelae at a later stage in life. 30° 30° National serosurveys for antibodies and antigenic mark- Re ers for carrier states of HBV are available for almost all d 20° Sea 20° nations at various stages of resolutions due to blood- banking practices. The overall carriage rate for Sub-Saharan 10° 10° Africa is 12 percent. These data allow for estimations of dis- ease burden based on a certain percentage of chronically 0° 0° infected persons progressing to hepatoma, fulminant hepa- ATLANTIC OCEAN INDIAN titis, or cirrhosis at later stages of life (Miller and McCann OCEAN 2000). For Sub-Saharan Africa, this would translate into 10° 10° more than 500,000 deaths per year in each birth cohort at a mean age of 40 to 45 years. As life expectancy in Sub- 20° 20° Saharan Africa is currently declining from the HIV epi- no demic, actual death from hepatitis is likely to be lower. yes 30° 30° no data Three doses of HBV vaccine, the first given at birth, could international boundaries This map was produced by the Map Design Unit of The World Bank. effectively prevent a child from becoming infected and The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. from becoming a chronic carrier of the virus. Vaccine ef- 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° fectiveness is approximately 95 percent with three doses. Source: WHO 2003. Therefore, HBV vaccine could exert a powerful herd effect on the population by eliminating the long latency period during which persons could be infectious to others. Despite mental retardation in patients who recover from the acute the benefits of implementing HBV into national vaccination form of the disease. campaigns, a minority of African nations are doing so As it is difficult to culture this organism, disease burden (figure 12.6). The WHO endorses several immunization estimates are made through a variety of modeling studies schedules so that HBV vaccine can be administered with integrating data from geographic representational popula- other vaccines. The vaccine does not have any known seri- tions and similar socioeconomic stratification (Miller 1998; ous effects and can be combined with other vaccines, Miller and McCann 2000). In a Hib vaccine trial in The including DTP and Hib. Gambia (Mulholland et al. 1997), the vaccinated group had As HBV is transmitted in a way similar to that of HIV, the 21 percent fewer occurrences of severe pneumonia than uptake and use of this vaccine is a good proxy for the poten- the control group, indicating that up to 21 percent of cases tial use of vaccines against HIV and the acquired immune of severe pneumonia may be due to this organism. deficiency syndrome (AIDS). Extrapolation to the rest of the region would indicate that this vaccine could potentially prevent 90,000 to 123,000 deaths due to bacterial pneumonia or meningitis. Haemophilus Influenzae type B Since 1988, safe and effective vaccines have been devel- Hib is a bacterium transmitted through respiratory excre- oped to prevent Hib. A conjugate vaccine can be coadminis- tions and may be carried in the nasopharynx of about tered with DTP, IPV, and HBV vaccine. A full course of 15 percent of nonimmunized children (WHO 1998). The vaccine confers more than 95 percent protection against most common forms of invasive disease are meningitis, invasive Hib disease and results in a herd effect (Wenger and epiglottitis, pneumonia, arthritis, and cellulitis. Epiglottitis Ward 2004). This vaccine may be given in combination with is characterized by swelling of the epiglottis, the tissue in the DTP. In countries that have introduced a three- to four-dose throat covering the larynx. The disease burden of Hib is regimen, Hib disease has almost been eliminated. The rela- highest among children between four and eighteen months, tively high cost of the Hib vaccine has hindered many low- rarely occurring in infants younger than three months and income countries from integrating the vaccine into routine after the age of six years. Hib meningitis could cause severe vaccine schedules. Research has been conducted to explain Vaccine-Preventable Diseases | 171 quantitatively the benefits of adding Hib to national immu- four doses of polysaccharide vaccine given during the first nization programs in order to assist policy makers in five years of life. Others (WHO 1997) have suggested rou- endemic countries. It has been estimated that from 257,000 tine immunization of schoolchildren or improvement of the to 317,000 deaths could be averted each year through the implementation of the current strategy of emergency routine use of Hib vaccine (Miller and McCann 2000). response with mass immunization. To date, these vaccines Although Hib vaccination would be a cost-effective inter- are not used routinely in Sub-Saharan Africa. A conjugate vention in low-income countries, its use in Africa has been meningococcal serogroup A vaccine currently being devel- limited. oped could routinely be administered to infants in Africa (LaForce 2004). Meningococcal Disease Neisseria meningitides is a bacterium that causes menin- FUTURE VACCINES gitis, frequently with devastating neurological sequelae (Heymann et al. 1998; Merlin et al. 1996) and a death rate Parallel to the successful delivery of vaccines, great advances often exceeding 30 percent, even with optimal treatment have been made in the field of vaccine development. Since (Campagne et al. 1999). Infections have also resulted in limb the 1985 U.S. Institute of Medicine report established prior- loss. Given its acute sporadic onset and devastating impact ities for vaccine development (IOM 1986), many new vac- on young healthy individuals, mostly less than 30 years of cines with the potential to dramatically reduce morbidity age, meningococcal disease is one of the most feared infec- and mortality have been developed. Within the last 20 years tious diseases. Although meningococcal disease is rare in new safe and effective vaccines have been developed and infants up to three months of age, the incidence gradually licensed to protect populations against HBV and Hib. Great increases to a peak at about one year of age and declines research strides in immunology have begun to pay divi- thereafter. dends with the more recent licensure of rotavirus vaccines Although there are five serogroups (A, B, C, Y, and against diarrhea, human papillomavirus vaccine against W135), serogroup A has caused annual epidemics in the genital warts and possible cervical cancer, and Streptococcus African meningitis belt, an area stretching from Senegal to pneumoniae (SP) vaccine against pneumonia. Although Ethiopia (Greenwood 1987; Greenwood et al. 1984; these vaccines are currently licensed, they are presently not Lapeyssonnie 1963). Annual incidence rates in this portion routinely used in Africa. of the African continent have exceeded 1,000 cases per 100,000 population in some instances (Granoff, Feavers, and Borrow 2004). In 1996 the largest recorded meningitis VACCINATION PROGRAMS outbreak occurred in Africa, resulting in more than 250,000 cases and 25,000 deaths (Tikhomirov, Santamaria, and The WHO objectives of the EPI in Africa are to strengthen Esteves 1997). the delivery of sustainable immunization services and to A polysaccharide vaccine containing polysaccharides accelerate efforts to achieve polio eradication, measles from serotypes A, C, Y, and W135 that is approximately control, neonatal tetanus elimination, and yellow fever con- 85 percent effective has been licensed but has been used only trol. Country strategic plans are prepared for the district in response to outbreaks triggered by a weekly incidence of level, with the goal of administering three doses of DTP to 10 to 15 cases per 100,000 persons in a geographically at least 80 percent of the target population in all districts defined region. Frequently, surveillance and logistical diffi- and ensure increasing government funding for the EPI. culties to timely identification of outbreaks and delivery of EPI programs are strengthening health systems throughout vaccines in these instances have had a limited impact on epi- Africa as well as implementing the use of new vaccines and demics. The 1996 outbreak and the continuous suboptimal technologies. implementation of emergency immunization campaigns Prior to the implementation of immunization programs, led to a reevaluation of meningococcal disease-prevention vaccine-preventable diseases were highly endemic through- strategies in this area (Miller and Shahab 2005). Robbins out Sub-Saharan Africa. During the 1990s the global immu- and colleagues (1997) called for immediate mass immu- nization coverage exceeded 70 percent (figure 12.1); how- nization, followed by a routine immunization program of ever, reported rates between regions and nations had great 172 | Mark A. Miller and John T. Sentz disparities. In Sub-Saharan Africa, reported immunization believed to be unsafe and pose the risk of transmitting rates peaked in 1990 at 55 percent and remained steady hepatitis, HIV, and other blood-borne pathogens (Miller throughout the decade. Immunization rates have decreased and Pisani 1999). In many circumstances, especially in over the last 10 years in many low-income countries, partic- developing countries, disposable syringes are used multiple ularly in Sub-Saharan Africa. For example, vaccination rates times, which increases the risk of disease transmission. of all three doses of DTP (DTP3) in the Central African Immunization rates drop dramatically when unsafe injec- Republic decreased from 82 percent in 1990 to 29 percent in tion practices are publicized, whether they even are a part of 2000. Similarly, the coverage rates with three doses of DTP vaccine campaigns. vaccine dropped in the Democratic Republic of Congo from Recognizing these factors, the WHO has recently been 79 percent in 1990 to 33 percent in 2000. These declines promoting the expansion and performance monitoring of have been attributed to civil unrest and lack of political services to local levels, with the goal of reaching routine cov- will among the national governments. As a result, millions erage of at least 80 percent. This may rectify the maldistribu- of children have been left unvaccinated and vulnerable to tion of vaccine delivery services within countries to increase disease. the equity of benefits, especially to the populations on the The success of vaccination programs applied at the global margins, those likely to have the greatest disease burden. level has had few parallels in public health. Immunizations remain one of the most cost-effective health interventions to prevent death and disability caused by infectious diseases. ADOPTION OF NEW VACCINES Despite great strides forward in vaccination development and administration throughout parts of the world, many The traditional EPI-targeted program has been challenged countries, usually the poorest, struggle with vaccinating by the development of new vaccines that promise to reduce their children. This gap in immunization coverage results the disease burden further if they can be adopted into from many compounding problems, such as low political national programs. Since the development of national commitment on behalf of national and local governments, vaccine programs, there has been a widening gap in the weak health service delivery systems, civil unrest, and number of vaccines used in developing and developed underfunding and poor management. These problems are countries (figure 12.7). further compounded by relatively low levels of research and development of new vaccines to combat the predominant diseases in the developing world. Figure 12.7 Number of Childhood Vaccines Routinely Used in There are many possible reasons why routine vaccination Developing and Established-Market Countries has declined in Africa. Over the past decade, the budget of the WHO vaccine program has been heavily skewed toward 14 the Poliomyelitis Eradication Initiative. Although this has 13 12 Influenza led to high coverage rates against poliomyelitis virus, it likely 11 Pneumococcal Meningococcal has strained other vaccination services (Taylor, Cutts, and 10 C Taylor 1997). Additionally, throughout the 1990s, donors 9 Varicella Haemophilus and the United Nations Children's Fund (UNICEF) empha- 8 Haemophilus influenzae influenzae Hepatitis Ba 7 Hepatitis B sized that countries should pay for vaccine out of their own vaccines Mumps Yellow fever of 6 Rubella national budget rather than finance the cost through no. Measles 5 DTP donors. The push toward self-reliance was made with good 4 Poliomyelitis BCG intentions, but fixed budgets within national economies 3 may have simply moved funds from vaccine administration 2 1 to vaccine purchase, with resultant declines in services. 0 An additional deterrent to vaccination adherence in 1975 1980 1985 1990 1995 2000 much of the world is the threat of unsafe injections. It has established market developing countries been estimated that nearly 8 billion to 12 billion injections are administered in various health care settings throughout Source: WHO and UNICEF 2003. the world each year; 50 percent of these injections are a. Used in about 50 percent of global birth cohort. Vaccine-Preventable Diseases | 173 The use of these vaccines within already existing vaccine Table 12.2 Potential Deaths Averted by HBV, Hib, Rotavirus, delivery infrastructures would greatly increase the arma- and SP Vaccine Implementation mentarium against infectious and even cancer-causing Estimated deaths Estimated program agents. However, despite the availability of these public Vaccine averted (thousands) costs (US$ millions) health tools, countries have been slow to adopt them. For HBV 978­1,370 671 example, despite the World Health Assembly's recommen- Hib 257­317 1,129 dation to add HBV vaccine to all member nations' vaccina- Rotavirus 149­326 1,482 tion programs by 1997 (WHA 1992), less than one-half the SP 549­1,098 3,729 global infant birth cohort receives this vaccine. Likewise, Source: Miller and McCann 2000. many African countries have yet to adopt the highly effica- cious Hib vaccine. Given the array of new vaccines on the horizon, and potential vaccines against HIV, malaria, and and McCann 2000). Disease burden, vaccine program costs, tuberculosis, what is the likelihood of their adoption, espe- and the potential reduction of disease from vaccination cially in those countries that have the highest mortality from were assessed for four vaccines that have not been adopted those diseases? in many countries. Factors such as national infrastructure to deliver vaccines Without vaccination, HBV, Hib, SP, and rotavirus con- and the cost of novel vaccines relative to national income in tribute to more than 1 million deaths in each successive addition to an inadequate appreciation of the disease bur- birth cohort in Africa (Miller and McCann 2000). Routine den have prevented the adoption of newer vaccines. The scheduled use of HBV, Hib, SP, and rotavirus vaccines could Global Alliance for Vaccines and Immunizations (GAVI) has potentially prevent most of these deaths (table 12.2). Incor- sought to address some of these factors through the cre- poration of these vaccines into routine vaccination pro- ation of vaccine purchase funds and the direct allocation of grams was estimated to cost between US$29 and US$150 financial resources to strengthen infrastructures at the local per life year saved (Miller and McCann 2000). Based on level. It is still too early to judge the success of these tactics these evaluations, HBV and Hib should be considered for and sustainability in the most resource-constrained areas integration into all national immunization programs. SP such as Africa. Development agencies will have limited and rotavirus vaccines, with the given assumptions, would influence on per capita gross domestic product, but they can also be cost-effective. Proactive analysis of the epidemio- work with manufacturers to influence the vaccine cost to the logic and economic impact of these vaccines can hasten country (either by offering financial support for vaccine their introduction into national vaccination schedules. purchase or guaranteeing volume purchase to lower vaccine prices) and strengthen vaccination services. FINANCIAL IMPLICATIONS, GAVI, AND THE VACCINE FUND VACCINATION: PROBLEMS AND BARRIERS GAVI is a public-private partnership devoted to the promo- The barriers to vaccine adoption are presumed to be multi- tion and strengthening of vaccine programs in low-income ple. If there is poor recognition of disease burden and costs, countries. Through the Vaccine Fund, which is the funding insufficient finances, or an ineffective vaccine delivery arm of GAVI, financial resources are provided for eligible system, the introduction of a new vaccine is unlikely. How- countries to purchase vaccines and to fund the operational ever, methodological analyses can help focus deliberations costs of managing immunization campaigns. GAVI provides and assist countries to overcome hurdles such as these. funding to national governments based on national income Demonstration of a country's disease burden as well as its (countries with gross national income per capita below associated economic burden may encourage the appropriate US$1,000). Approximately half of the 75 countries eligible allocation of financial resources by clarifying the value of for funding through GAVI are located in Sub-Saharan prevention through vaccination. Africa. In order for an eligible country to receive funding Economics has frequently been sited as a barrier to the it must demonstrate that a well-functioning national mech- use of newer vaccines. To address this, an analysis of the anism is in place to coordinate immunization activities health and economic implications of new vaccine introduc- among the various partners within the country. The immu- tion was conducted to help national policy makers (Miller nization program must have received a comprehensive 174 | Mark A. Miller and John T. Sentz assessment by GAVI-designated technical agencies during In many parts of Africa, vaccine infrastructure has been the previous three years, a multiyear immunization plan suboptimal, especially for routine vaccination. Immuniza- must be complete, and finally a strategy for improving the tion campaigns versus routine services appear to be domi- safety of injections must be demonstrated. Currently the nating given logistical and operational hurdles in Sub- GAVI's Vaccine Fund is providing financial support to gov- Saharan Africa ernments in order to improve health services; distribute safe In Africa, new vaccines have been slow to be adopted into injection materials; and acquire the HBV, Hib, and yellow national EPI programs; likely reasons have been the lack of fever vaccines. support for routine delivery throughout the continent. In an attempt to ensure the sustainability of immu- Logistical and operational factors are probable barriers that nization campaigns, GAVI requires countries that receive require substantial continuous investments in human capi- funding to create financial sustainability plans. Financial tal, equipment, and financing. sustainability of immunization campaigns is vital in order to reach any long-term immunization goals. These financial sustainability plans are meant to be long-term plans that can REFERENCES improve the countries' immunization programs and serve as a foundation to expand support from other sources, namely, Anita-Obong, O. E., M. U. Young, and C. E. Effiong. 1993. "Neonatal their own national budgets. Tetanus: Prevalence before and Subsequent to Implementation of the Expanded Programme on Immunization." Annals of Tropical Initiatives by bilateral partner and health agencies empha- Paediatrics 13: 7­12. size the need for nations to strengthen health system infra- Babaniyi, O., and B. Parakoyi. 1991. "Cluster Survey for Poliomyelitis structures and assume responsibility for the provision of basic and Neonatal Tetanus in Ilorin, Nigeria." International Journal of Epidemiology 20 (2): 515­20. services, including childhood immunizations. Concurrent Black, F. L. 1976. "Measles." In Viral Infections of Humans: Epidemiology with these structural changes within nations are pricing and and Control, ed. A. S. Evans. New York: Plenum. technology changes in vaccines. The six basic EPI vaccines are Campagne, G., A. Schuchat, S. Djibo, A. Ouseini, L. Cissé, and J. P. currently estimated to cost approximately US$0.50 per capita. Chippaux. 1999. "Epidemiology of Bacterial Meningitis in Niamey, Niger, 1981­96." Bulletin of the World Health Organization 77: 499­508. New vaccines are appreciably more costly than traditional Cantwell, M. F., and N. J. Binkin. 1996. "Tuberculosis in Sub-Saharan vaccines but far less expensive than the treatment costs of the Africa: A Regional Assessment of the Impact of the Human diseases they prevent. The poorer developing countries, Immunodeficiency Virus and National Tuberculosis Control Program Quality." Tubercle and Lung Disease 77 (3): 220­25. which often have the greatest disease burden, may be inclined Cherry, J. D., and R. E. Harrison. 2004. "Tetanus." In Textbook of Pediatric to reject the adoption of these vaccines without full recogni- Infectious Diseases, ed. R. D. Feigin, J. D. Cherry, G. J. Demmler, and tion of their true value. S. L. Kaplan. Philadelphia: Elsevier. A determination of the value of prevented disease is not CVI (Children's Vaccine Initiative) Forum 08. 1994. Neonatal Tetanus: The Final Countdown. Geneva: WHO. a panacea for a shortage of financial resources. However, it De Serres, G., R. Shadmani, B. Duval, N. Boulianne, P. Dery, M. Fradet, can support the more efficient allocation of limited L. Rochette, and S. Halperin. 2000. "Morbidity of Pertussis in resources. Indeed, many new financial resources, most Adolescents and Adults." Journal of Infectious Diseases 182: 174­79. notably from the World Bank and the Bill & Melinda Gates Dragsted, D., B. Dohn, J. Madsen, and J. Jenson. 2004. "Comparison of Culture and PCR for Detection of Bordetella Pertussis and Bordetella Foundation, are expanding the number of potential Parapertussis under Routine Laboratory Conditions." Journal of resources available to help finance vaccination activities. Medical Microbiology 53: 749­54. There is concurrently an expanded role for medical associ- Edwards, K. M., and M. D. Decker. 2004. "Pertussis Vaccine." In Vaccines, 4th ed., ed. S. A. Plotkin and W. A. Orenstein. Philadelphia: Saunders. ations on the national, regional, and global level to highlight Elmore-Meegan, M., R. M. Conroy, S. Ole Lengeny, K. Renhault, and the value of prevention through use of some of the best J. Nyangole. 2001. "Effect on Neonatal Tetanus Mortality after a tools in public health. Culturally-Based Health Promotion Programme." Lancet 358: 640­41. Feigin, R. D., B. W. Stechenberg, and P. Hertel. 2004. "Diptheria." In Textbook of Pediatric Infectious Diseases, ed. R. D. Feigin, J. D. Cherry, G. J. Demmler, and S. L. Kaplan, 1305­13. Philadelphia: Elsevier. CONCLUSION Fine, P. E. M. 2001. "BCG Vaccines and Vaccination." In Tuberculosis: A Comprehensive International Approach, ed. L. B. Reichman and E. S. Traditional vaccines are among the most cost-effective inter- Hershfield. New York: Marcel Dekker. ventions there are and can potentially be used to achieve Fry, N., O. Tzivra, T. Li, A. McNiff, N. Doshi, C. Maple, N. Crowcroft, E. Miller, R. George, and T. Harrison. 2004. "Laboratory Diagnosis of several Millennium Development Goals; however, continu- Pertussis Infections: The Role of PCR and Serology." Journal of Medical ous vaccination must be sustainable. Microbiolgy 53: 519­25. Vaccine-Preventable Diseases | 175 Galazka, A. 1992. "Control of Pertussis in the World." World Health Miller, E., and N. J. Gay. 1997. "Epidemiological Determinants of Statistics Quarterly 45: 238­47. Pertussis." Developments in Biological Standardization 89: 15­23. Galazka, A., and S. Robertson. 1996. "Immunization against Diphtheria Miller, M., S. Barrett, and D. A. Henderson. 2006. "Control and with Special Emphasis on Immunization of Adults." Vaccine 14 (9): Eradication." In Disease Control Priorities in Developing Countries, 845­57. 2nd ed, ed. D. T. Jamison, J. G. Breman, A. R. Measham, G. Alleyne, Galazka, A., S. Robertson, and G. Oblapenko. 1995. "Resurgence of M. Claeson, D. B. Evans, P. Jha, A. Mills, and P. Musgrove. New York: Diphtheria." European Journal of Epidemiology 11: 95­105. Oxford University Press. Geldermalsen, A. A., and U. Wenning. 1993. "A Diphtheria Epidemic in Miller, M. A. 1998. "An Assessment of the Value of Haemophilus Lesotho, 1989. Did Vaccination Increase the Population's Influenzae Type B Conjugate Vaccine in Asia." Pediatric Infectious Susceptibility?" Annals of Tropical Paediatrics. 13: 13­20. Disease Journal 17 (Suppl. 9): S152­59. Gelfand, H. M., D. R. LeBlanc, J. P. Fox, and D. P. Conwell. 1957. "Studies ------. 2000. "Introducing a Deterministic Model to Estimate Global on the Development of Natural Immunity to Poliomyelitis in Measles Disease Burden." Journal of International Infectious Diseases 4: Louisiana. II. Description and Analysis of Episodes of Infection 14­20. Observed in Study Households." American Journal of Hygiene 65: Miller, M. A., and L. McCann. 2000. "Policy Analysis of the Use of 367­85. Hepatitis B, Haemophilus Influenzae Type B, Streptococcus Granoff, D. M., I. M. Feavers, and R. Borrow. 2004. "Meningococcal Pneumoniae-Conjugate and Rotavirus Vaccines in National Vaccines." In Vaccines, ed. S. A. Plotkin and W. A. Orenstein. Immunization Schedules." Health Economics 9: 19­35. Philadelphia: Saunders. Miller, M. A., and E. Pisani. 1999. "The Cost of Unsafe Injections." Bulletin Granstrom, G., B. Wretlind, and M. Granstrom. 1991. "Diagnostic Value of of the World Health Organization 77 (10): 808­11. Clinical and Bacteriological Findings in Pertussis." Journal of Infection Miller, M. A., and C. K. Shahab. 2005. "Review of the Cost Effectiveness 22: 17­26. of Immunization Strategies for the Control of Meningococcal Greenwood, B. M. 1987. "Epidemiology of Meningitis in Tropical Africa." Meningitis." Pharmacoeconomics 23 (4): 333­43. In Bacterial Meningitis, 1st ed., ed. J. D. Williams and J. Burnie, 61­91. Monath, T. 2004. "Yellow Fever Vaccine." In Vaccines, ed. S. A. Plotkin and London: Academic Press. W. A. Orenstein. Philadelphia: Saunders. Greenwood, B. M., I. S. Blakebrough, A. K. Bradley, S. Wali, and H. C. Mulholland, K., S. Hilton, R. Adegbola, S. Usen, A. Oparaugo, C. Whittle. 1984. "Meningococcal Disease and Season in Sub-Saharan Omosigho, M. Weber, et al. 1997. "Randomised Trial of Haemophilus Africa." Lancet 1: 1339­42. influenzae Type-b Tetanus Protein Conjugate Vaccine [Corrected] for Heyman, S. N., Y. Ginosar, L. Niel, J. Amir, N. Marx, M. Shapiro, and S. Prevention of Pneumonia and Meningitis in Gambian Infants." Lancet Maayan. 1998. "Meningococcal Meningitis among Rwandan Refugees: 349: 1191­97. Diagnosis, Management, and Outcome in a Field Hospital." Patriarca, P. A., R. J. Biellik, G. Sanden, D. G. Burstyn, P. D. Mitchell, P. R. International Journal of Infectious Diseases 2: 137­42. Silverman, J. P. Davis, and C. R. Manclark. 1988. "Sensitivity and Heymann, D. L., and R. B. Aylward. 2004."Eradicating Polio." New England Specificity of Clinical Case Definitions for Pertussis." American Journal Journal of Medicine 351 (13): 1275­77. of Public Health 78: 833­36. Heyworth, B., and M. Ropp. 1973. "Diphtheria in The Gambia." Journal of Plotkin, S., and E. Vidor. 2004. "Poliovirus Vaccine-Inactivated." In Tropical Medicine and Hygiene 76 (3): 61­64. Vaccines, ed. S. A. Plotkin, and W. A. Orenstein. Philadelphia: Saunders. Horstmann, D. M. 1955. "Poliomyelitis: Severity and Type of Disease in Postels-Multani, S., H. J. Schmitt, C. H. Wirsing von Konig, H. L. Bock, Different Age Groups." Annals of the New York Academy of Sciences 61: and H. Bogaerts. 1995. "Symptoms and Complications of Pertussis 956­67. in Adults." Infection 23: 139­42. Idema, C. D., B. N. Harris, G. A. Ogunbanjo, and D. N. Durrheim. 2002. Preziosi, M. P., A. Yam, S. G. Wassilak, L. Chabirand, A. Simaga, M. Ndiaye, "Neonatal Tetanus Elimination in Mpumalanga Province, South M. Dia, et al. 2002. "Epidemiology of Pertussis in a West African Com- Africa." Tropical Medicine and International Health 7 (7): 622­24. munity before and after Introduction of a Widespread Vaccination Program." American Journal of Epidemiology 155 (10): 891­96. IOM (Institute of Medicine). 1986. New Vaccine Development: Diseases of Importance in Developing Countries. Vol. 2. Washington, DC: National Robbins, J., D. W. Towne, E. C. Gotschlich, and R. Schneerson. 1997. Academy Press. "Love's Labours Lost: Failure to Implement Mass Vaccination against Group A Meningococcal Meningitis in Sub-Saharan Africa." Lancet Kew, O. M., P. F. Wright, V. I. Agol, F. Delpeyroux, H. Shimizu, N. 350: 880­82. Nathanson, and M. A. Pallansch. 2004. "Circulating Vaccine-Derived Polioviruses: Current State of Knowledge." Bulletin of the World Health Rodrigues, L. 1991. "EPI Target Diseases: Measles, Tetanus, Polio, Organization 82: 16­23. Tuberculosis, Pertussis, and Diphtheria." In Disease and Mortality in Sub-Saharan Africa, ed. R. Feachem and D. Jamison, 173­89. New LaForce, F. M. 2004. "Conjugate Meningococcal Vaccines for Africa." In York: Oxford University Press. Vaccines: Preventing Disease and Protecting Health, ed. C. A. de Quadros. Washington, DC: PAHO. Sabin, A. B. 1951. "Paralytic Consequences of Poliomyelitis Infection in Different Parts of the World and in Different Population Groups." Lapeyssonnie, L. 1963. "La méningite cérébrospinale en Afrique American Journal of Public Health 41: 1215­30. [Cerebrospinal meningitis in Africa]." Bulletin of the World Health Organization 28 (Suppl.): 1­100. Smith, K. C., and J. Starke. 2004. "Baccille Calmette-Guérin Vaccine." In Vaccines, ed. S. A. Plotkin, and W. A. Orenstein. Philadelphia: Saunders. Mahieu, J. M., A. S. Muller, A. M. Voorhoeve, and H. Dikken. 1978. "Pertussis in a Rural Area of Kenya: Epidemiology and a Preliminary Stanfield, J. P., and A. Galazka. 1984. "Neonatal Tetanus in the World Report on a Vaccine Trial." Bulletin of the World Health Organization 56 Today." Bulletin of the World Health Organization 62: 647­49. (5): 773­80. Strebel, P., M. Papania, and N. Halsey. 2004. "Measles Vaccine." In Vaccines, Merlin, M., G. Martet, J. M. Debonne, P. Nicolas, C. Bailly, D. Yazipo, J. ed. S. A. Plotkin, and W. A. Orenstein. Philadelphia: Saunders. Bougere, A. Todesco, and R. Laroche. 1996. "Controle d'une epidemie Sutter, R. W., and O. M. Kew. 2004. "Poliovirus Vaccine-Live." In Vaccines, de meningite a meningocoque en Afrique centrale [Control of an epi- ed. S. A. Plotkin, and W. A. Orenstein, 651­705. Philadelphia: demic of meningococcal meningitis in Central Africa]." Santé 6: 87­95. Saunders. 176 | Mark A. Miller and John T. Sentz Taylor, C. E., F. Cutts, and M. E. Taylor. 1997."Ethical Dilemmas in Current ------. 1999b. "Pertussis Vaccines." Weekly Epidemiological Record 74: Planning for Polio Eradication." American Journal of Public Health 87 137­44. (6): 922­25. ------. 1999c. "Progress towards the Global Elimination of Neonatal Tikhomirov, E., M. Santamaria, and K. Esteves. 1997. "Meningococcal Tetanus, 1990­1998." Weekly Epidemiological Record 74: 73­80. Disease: Public Health Burden and Control." World Health Statistics ------. 2001a. Global Tuberculosis Control: WHO Report 2001. Quarterly 50: 170­77. WHO/CDS/TB/2001.287. Geneva: WHO. Wenger, J., and J. Ward. 2004. "Haemophilus influenza Vaccine." In ------. 2001b. Maternal and Neonatal Tetanus Elimination: African Vaccines, ed. S. A. Plotkin and W. A. Orenstein, 229­68. Philadelphia: Region. WHO/UNICEF Five Year Regional Plan of Action, 2001­2005. Saunders. Geneva: WHO. Wharton, M., and C. Vitek. 2004. "Diptheria Toxoid." In Vaccines, ed. S. A. ------. 2003. WHO Vaccine-Preventable Disease Monitoring System: 2003 Plotkin and W. A. Orenstein, 211­28. Philadelphia: Saunders. Global Summary. WHO/V&B.03.20. Geneva: WHO. WHA (World Health Assembly). 1992. "Immunization and Vaccine ------. 2004. The World Health Report 2004--Changing History. Geneva: Quality." Forty-Fifth World Health Assembly, Geneva, May 4­14. WHO. Resolution WHA45.17, World Health Assembly, Geneva. ------. 2006. "Conclusions and Recommendations from the WHO (World Health Organization). 1985. "Report of the Expanded Immunization Strategic Advisory Group." Weekly Epidemiological Programme on Immunization Global Advisory Group Meeting." Record 81: 2­11. WHO/EPI/GEN/87/1, WHO, Geneva. WHO (World Health Organization) and UNICEF (United Nations ------. 1997. "Minutes." WHO Informal Consultation on Operational Children's Fund). 1996. State of the World's Vaccines and Immunization. Research on Immunization Prevention and Control Strategies for Geneva: WHO. Meningococcal Disease, Annecy, France, February 18­19. WHO/ EMC.97.2, WHO, Geneva. ------. 2003. State of the World's Vaccines and Immunization. Geneva: ------. 1998. "Global Programme for Vaccines Immunization." Position WHO. Paper on Haemophilus influenzae type B conjugate vaccines. Weekly World Bank. 1993. World Development Report 1993: Investing in Health. Epidemiological Record 73: 64­68. New York: Oxford University Press. ------. 1999a. Field Manual for Neonatal Tetanus Elimination. Geneva: ------. 2003. World Development Indicators. Washington, DC: World WHO. Bank. Vaccine-Preventable Diseases | 177 Chapter 13 Tuberculosis Christopher Dye, Anthony D. Harries, Dermot Maher, S. Mehran Hosseini, Wilfred Nkhoma, and Felix M. Salaniponi About 9 million people around the world developed tuber- parts of the continent (Cauthen, Pio, and ten Dam 2002). culosis (TB) for the first time in 2004, and nearly 2 million The burden of TB in Sub-Saharan Africa is far greater today. people died with or from the disease. Globally, TB is cur- Continuing poverty and political instability in parts of the rently responsible for more years of healthy life lost (2.5 per- continent has inhibited progress in implementing effective cent of all disability-adjusted life years, or DALYs) than any TB control measures. But the principal reason for the resur- other infectious disease, bar AIDS and malaria (Corbett gence of TB in Africa is not the deterioration of control et al. 2003; WHO 2002; WHO 2006). Only AIDS is respon- programs. Rather, it is the link between TB and the human sible for more deaths. The full cost of the worldwide TB immunodeficiency virus and the acquired immune defi- epidemic is rarely appreciated. The direct monetary costs of ciency syndrome (HIV/AIDS). People who are latently diagnosis and treatment are borne by health services and by infected with Mycobacterium tuberculosis--about one-third patients and their families. Added to these are the indirect of the inhabitants of Sub-Saharan Africa (Dye et al. 1999)-- costs of lost income and production, incurred when TB are at hugely greater risk of developing active TB if they are patients are too sick to work and when young adults--often also immunologically weakened by a concurrent HIV infec- parents and householders--die prematurely (WHO 2000). tion. HIV-positive people are also more likely to develop TB Beyond these losses, baldly expressed in DALYs and dollars, when newly infected or reinfected with M. tuberculosis. Over enormous psychological and social costs are associated with the past decade, the TB caseload has increased by a factor of TB. These extra costs are less easily quantified, but they are five or more in those countries of eastern and southern nonetheless real. Africa that are most affected by HIV. Incidence rates in A decade ago the problem of TB in Africa attracted little these countries are now comparable with those recorded attention, not even meriting a chapter in the first edition of in Europe half a century ago, before the introduction of Disease and Mortality in Sub-Saharan Africa. Part of the antituberculosis drugs. reason was that TB incidence was low and falling in most 179 MICROBIOLOGY, TRANSMISSION, made use of in culture techniques that use alkalis and special AND PATHOGENESIS media to reduce contamination. M. tuberculosis bacilli transmitted on airborne droplets cause, most importantly, a lung disease that will kill about Transmission and Risk of Infection half of all untreated patients. People infected with M. tuberculosis carry live tubercle bacilli, but the bacilli may be present in small numbers and dormant (latent), in which case there may be no apparent Microbiology disease. Disease occurs when the bacteria multiply, over- The M. tuberculosis complex includes five species: M. tuber- come immune defenses, and become numerous enough to culosis, M. bovis (and bacillus Calmette-Guérin), M. canetti, cause damage to tissues. M. africanum, and M. microti. Within the species complex, Patients with pulmonary tuberculosis (PTB) are the most most human disease is due to M. tuberculosis sensu stricto. important source of infection. Infection occurs by inhaling The variants within the species complex differ from the type droplet nuclei, infectious particles of respiratory secretions strain biochemically and in culture. However, these differ- usually less than 5 micrometers, which contain tubercle ences have no known bearing on management or prognosis. bacilli. These are spread into the air by coughing, sneezing, The principal exception is M. bovis, which accounts for a talking, spitting, and singing, and they can remain suspended small fraction of human TB cases, but which is naturally in the air for long periods of time. A single cough can resistant to the drug pyrazinamide (which should not there- produce 3,000 infectious droplet nuclei. Direct sunlight kills fore be used in treatment). tubercle bacilli in minutes, but they can survive in dark, Human disease can also be caused by species of mycobac- unventilated environments for longer periods of time. teria other than M. tuberculosis (MOTT), also known as Droplet nuclei are so small that they avoid the defenses of atypical mycobacteria. These organisms are widespread in the bronchi and penetrate into the terminal alveoli of the nature and have been isolated from a variety of sources, lungs, where multiplication and infection begins. including soil, dust, water, milk, animals, and birds. In The risk of infection is determined by the infectiousness humans, MOTT are low-grade pathogens and usually cause of the source case (that is, how many tubercle bacilli are disease only in patients with preexisting lung disease or being coughed into the air), the closeness of contact, light immunodeficiency. MOTT are still a rare cause of disease in and humidity, and the immune status of the host (Rieder Sub-Saharan Africa. The large majority of patients in Africa 1999). Patients with sputum smear-positive pulmonary who are diagnosed and treated for TB, even those infected disease (tubercle bacilli visible under the microscope when with HIV, have disease caused by M. tuberculosis (Nunn, appropriate stains are used) are much more infectious than Elliott, and McAdam 1994). those with smear-negative sputum (Styblo 1991). Following Mycobacteria are acid and alcohol fast, meaning that infection, the tubercle bacilli multiply in the lungs, spread to once stained by an aniline dye, such as carbolfuchsin, they the local lymph nodes, and then to the rest of the body. resist decolorization with acid and alcohol. Mycobacteria are About six weeks after this primary infection, the body devel- therefore often called "acid-fast bacilli" (AFB). In virtually ops an immune response to the tubercle bacilli called all other bacteria the dye is removed by the acid-alcohol delayed hypersensitivity. In the majority of cases, the wash, and the ability of mycobacteria to retain the aniline immune response stops the further multiplication of the dye despite acid and alcohol is probably due to their thick tubercle bacilli, and the only evidence of infection is a cell wall. This property allows the detection of AFB in positive response to an immunological test, of which the specimens by using the simple Ziehl-Neelsen (ZN) staining most commonly used is the tuberculin skin test (Ewer et al. technique, widely used in Sub-Saharan Africa. 2003; Mazurek and Villarino 2003; Von Pirquet 1909). Mycobacteria grow slowly, with generation times mea- The proportion of any population infected depends on sured in hours rather than minutes. This means that the the rate and duration of exposure, and this varies from one normal methods of obtaining cultures from clinical speci- group of people to another. There are, however, some mens are difficult because of overgrowth by other bacteria. common patterns. Because infection can remain dormant Fortunately, the thick cell wall of mycobacteria also enables for many years, or because the immunological consequences them to resist alkalis and detergents, and this property is of infection are long-lasting, infection rates are always 180 | Christopher Dye, Anthony D. Harries, Dermot Maher, S. Mehran Hosseini, Wilfred Nkhoma, and Felix M. Salaniponi observed to increase monotonically with age. Although the CLINICAL MANIFESTATIONS AND DIAGNOSIS infection rates in boys and girls are usually indistinguish- able, adult men typically show higher infection rates than Clinical diagnoses of tuberculosis distinguish between adult women. Men probably suffer more from TB than pulmonary and extrapulmonary disease, the former being women, not because they are more susceptible to disease, of much greater importance epidemiologically. but because they are more exposed to infection. Pulmonary Tuberculosis Pathogenesis Patients with pulmonary TB present with a chronic produc- The size of the infecting dose of tubercle bacilli and the tive cough, fever, and weight loss. Cough occurs in a variety immune status of the host determine the risk of progression of circumstances, notably in acute upper and lower respira- from infection to disease. When infection progresses to tory infections. However, these acute infections often resolve disease, it is manifest as infiltrates and lesions within the within three weeks. Therefore a patient with a cough longer lung tissue, enlarged lymph nodes within the chest, pleural than three weeks, which persists after a course of antibiotics, effusion, or disease disseminated in other parts of the body. should be investigated for PTB. The immune response of the patient results in a pathological The diagnosis of PTB in most hospitals in African lesion, which is characteristically localized, often with ex- countries is based on sputum smear microscopy and chest tensive tissue destruction and cavitation. These cavitating radiography. Most countries have a reference laboratory lesions occur most commonly in the lungs and contain where M. tuberculosis can be cultured from clinical speci- many actively dividing bacilli. Sputum from patients with mens, such as sputum. Because M. tuberculosis is a slow- these lesions is usually smear positive. growing organism taking two to three months to become If the primary infection resolves, small numbers of visible on culture medium, cultures are not usually helpful tubercle bacilli can remain dormant in scarred areas of the in making an individual diagnosis. Mycobacterial cultures body for many years. Postprimary TB may then occur by the are commonly used for monitoring drug-sensitivity pat- process of endogenous reactivation, and it may arise in any terns in patients with recurrent TB and for monitoring the other organ system to which the tubercle bacilli were seeded community prevalence of drug-resistant TB. during the primary infection. Active disease can also follow Taking sputum specimens (three per suspect) for smear from secondary or exogenous reinfection in a person who microscopy of AFB is a cheap and simple way to screen for already has a latent infection. PTB. However, sputum smears may be negative in pul- Without HIV coinfection, the average lifetime risk of monary TB patients for three reasons. First, the patient has infected individuals' developing tuberculosis is 5 to 10 per- genuine smear-negative pulmonary tuberculosis, that is, cent, the highest risk being within the first five years of expectorating small numbers of AFB. AFB can be detected infection (Comstock, Livesay, and Woolpert 1974; on microscopy only if there are 10,000 organisms or more Sutherland 1976).The risk of developing TB following infec- per milliliter of sputum. Second, the clinical diagnosis of tion also changes with age. Infants and young children up to TB is incorrect and the patient has another condition, such the age of five years who are infected with M. tuberculosis are as left ventricular failure, asthma, bacterial pneumonia, at relatively high risk, particularly of severe forms (mainly Pneumocystis carinii pneumonia, or pulmonary Kaposi's miliary TB and TB meningitis), because of their immature sarcoma. Third, the result is a false negative resulting immune systems. Children between the ages of five and from technical inadequacies (poor sputum sample, faulty fifteen years are relatively resistant to TB. The risk then rises smear preparation, inadequate time spent examining the again through adolescence, remains approximately stable smear) or administrative failures (incorrect labeling of during adulthood, but increases again in the elderly. specimens). Other factors that enhance the risk of developing TB If sputum smear examination shows no AFB, patients following infection include undernutrition, toxins (tobacco, suspected of having PTB should be referred for chest radi- alcohol, corticosteroids, immunosuppressive drugs), and ography. Classical patterns of TB with upper lobe disease, other diseases (diabetes mellitus, silicosis, leukemia, measles, bilateral disease, and cavitations are more common in HIV- and whooping cough in children), but none is as important negative patients and in HIV-positive patients who have as HIV (Crofton, Horne, and Miller 1999; Rieder 1999). relatively well preserved immune function. However, no Tuberculosis | 181 chest radiographic pattern is absolutely diagnostic of TB Given the problems with diagnosis and the low frequency (Hargreaves et al. 2001). of routine childhood screening, the real burden of child- hood TB in Sub-Saharan Africa is not known. One nation- wide study in Malawi found that 12 percent of all registered Extrapulmonary TB TB cases were children less than 15 years of age (Harries, Hargreaves, Graham et al. 2001). Such investigations are a The common signs or forms of extrapulmonary TB (EPTB) useful starting point for much-needed, population-based are pleural effusion, lymphadenopathy, pericardial effusion, studies of the burden of TB in children. miliary disease, and meningitis. Patients usually present with constitutional symptoms and local features related to the site of disease. EPTB is found in a higher proportion of Consequences of HIV Coinfection female TB patients than male patients. If patients cough for Untreated HIV infection causes a progressive decline in the longer than three weeks, sputum smear examination and number of CD4+ T lymphocytes and progressive dysfunc- chest radiography are often carried out, because patients tion of those lymphocytes that survive. CD4+ cells play a may have coexisting pulmonary disease. Definitive diagno- major role in the body's defense against tubercle bacilli, and sis of EPTB depends on having diagnostic tools, such as it is therefore not surprising that HIV infection is the most radiographs, ultrasound scans, procedures to obtain and powerful known risk factor for progression to active disease analyze fluid samples, and procedures for tissue biopsies and in those with a latent M. tuberculosis infection. In HIV- histological analysis. This degree of diagnostic sophistica- positive people infected with M. tuberculosis the annual risk tion is often unavailable in district hospitals in Africa. For of developing active disease is 5 to 15 percent, with a lifetime example, in one study in Tanzania only 18 percent of risk of 50 percent or higher (Lienhardt and Rodrigues 1997; patients diagnosed with EPTB had laboratory confirmation Raviglione et al. 1997; Rieder et al. 1989). HIV-infected of the diagnosis (Richter et al. 1991). people are also more susceptible to new tuberculous infec- tions (Di Perri et al. 1989) and to reinfection, and they progress more frequently and more quickly to overt disease Childhood TB (Sonnenberg et al. 2001). After the end of a primary episode, Children are most commonly infected with M. tuberculosis HIV increases the likelihood that TB will recur (Fitzgerald, as a result of transmission from an adult (often a family Desvarieux et al. 2000), either by reactivation (true relapse) member) with smear-positive disease. Most children remain or reinfection (Daley 1993). asymptomatic, and a positive tuberculin test may be the People who are coinfected with M. tuberculosis and HIV only evidence of infection. For those who do progress to can develop TB across a wide spectrum of immunodeficiency disease, PTB is the most common manifestation in both (Ackah et al. 1995; Mukadi et al. 1993), but the risk of devel- HIV-infected and HIV-uninfected children, although extra- oping active disease increases as the CD4+ cell count pulmonary disease is more frequent in those who are HIV declines (figure 13.1). HIV-positive patients with moderate positive. The patterns of EPTB in children and the diagnostic to severe immunosuppression show atypical forms of TB, problems encountered are similar to those described for which complicates radiographic diagnosis. Some HIV- adults, although meningitis makes up a higher proportion infected TB patients have normal chest radiographs and of EPTB cases in young children. negative sputum smears, and the diagnosis may therefore be The diagnosis of childhood PTB has always been difficult missed unless there are pronounced clinical signs or symp- because children rarely produce sputum for smear examina- toms. Among children, the highest rates of HIV infection tion. Diagnosis therefore usually requires a combination of are observed in those age one to four years, and the accurate clinical features, history of contact with a sputum-positive diagnosis of TB has become especially difficult in this age case, growth faltering, chest X-ray, and tuberculin skin test. group (Graham, Coulter, and Gilks 2001). In children and Chest X-ray findings on their own are nonspecific, as are adults who are infected with M. tuberculosis and who are clinical features, but the most important symptoms are also HIV positive, immunosuppression frequently leads to weight loss and poor appetite. Gastric aspiration, induced negative tuberculin skin tests. sputum, and nasopharyngeal aspiration show promise as The presentation of HIV-related EPTB is generally no alternative diagnostic techniques but are not practical under different from that of HIV-negative EPTB, although there routine clinical conditions in Africa. are sometimes complications. The enlargement of TB 182 | Christopher Dye, Anthony D. Harries, Dermot Maher, S. Mehran Hosseini, Wilfred Nkhoma, and Felix M. Salaniponi Figure 13.1 Relative Incidence of TB in HIV-Infected EPIDEMIOLOGY Individuals as a Function of CD4 Cell Count About one-third of the population of Sub-Saharan Africa is 0.02 infected with M. tuberculosis (Dye et al. 1999). In the year 2000, an estimated 17 million people in Sub-Saharan Africa TB of were infected with both M. tuberculosis and HIV--70 per- cent of all people co-infected worldwide (Corbett et al. 0.01 incidence 2003). As more people have become infected and coinfected with HIV, especially in eastern and southern Africa, the relative incidence of TB has been driven upward, as reflected in estimates derived from population-based surveys and from 0 200 400 600 800 routine TB surveillance data (figures 13.2 and 13.3) (WHO CD4 cell count/microliter 2002, 2006). In 2004, the incidence rate of TB in the WHO African region was growing at approximately 3 percent per Source: Williams and Dye 2003. Note: TB incidence is nominally set to 100 per 100,000 people at 800 CD4+ per microliter. The year (table 13.1), and at 4 percent per year in eastern and middle line is the best estimate, and the outer lines, 95 percent confidence limits (CL), were obtained from a combined analysis of data from three separate studies. southern Africa (the areas most affected by HIV), faster than on any other continent, and considerably faster than the lymph nodes in HIV-positive patients can occasionally be 1 percent per year global increase (WHO 2006). In several rapid and resemble an acute abscess. It is possible that a African countries, including those with well-organized diagnosis of miliary or disseminated TB is regularly missed. control programs (Harries et al. 1996; Kenyon et al. 1999), Diagnosis is often more difficult in patients who are severely annual TB case-notification rates have risen more than five- immunosuppressed. For example, disseminated TB was fold since the mid-1980s, reaching more than 400 cases per diagnosed only after death in 44 percent of patients with 100,000 people (WHO 2006). HIV infection is the most HIV wasting syndrome in Côte d'Ivoire; TB was not recog- important single predictor of TB incidence across the nized during life (Lucas et al. 1994). African continent (figure 13.4). Despite the emphasis placed Figure 13.2 Estimated Incidence Rates of New TB Cases and Percentages of TB Patients Infected with HIV, by Country, 2004 a. TB cases (all forms) per 100,000 people b. HIV prevalence in TB cases, age 15­49 years (percent) 30° 20° 10° 0° 10° 20° 30° 40° 50° 30° 20° 10° 0° 10° 20° 30° 40° IBRD 34528 MARCH 2006 Mediterranean Mediterranean Sea Sea 30° 30° 30° 30° Red Red 20° Sea 20° 20° Sea 20° 10° 10° 10° 10° 0° ATLANTIC 0° 0° ATLANTIC 0° OCEAN INDIAN OCEAN INDIAN OCEAN OCEAN 10° 10° 10° 10° less than 5 20° less than 100 20° 20° 20° 5 ­ 14 100 ­ 299 15 ­ 49 300 and more 50 and more 30° no data 30° 30° 30° no data international boundaries This map was produced by the Map Design Unit of The World Bank. international boundaries The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° Source: Data from WHO 2006. Tuberculosis | 183 Figure 13.3 Trends in TB (All Forms) Case Notifications in the Figure 13.4 Estimated TB Incidence in Relation to Estimated WHO African Region Contrasted with Trends in Other Parts of HIV Prevalence in Adults Age 15 to 49 for 42 Countries in the the World WHO African Region (per 100,000 people) 6 1,000 year 5 per 800 4 cases TB 3 600 incidence new TB of 2 400 1 millions estimated 200 0 1990 1995 2000 2005 year 0 10 20 30 40 % estimated HIV prevalence Asia Africa other regions Source: Corbett et al. 2003. Source: Data from WHO 2006. Note: Asia includes the WHO Southeast Asia and Western Pacific regions. Table 13.1 The Contribution of Sub-Saharan Africa to the Global TB Epidemic Estimated New Cases of TB HIV-related TB Deaths Increase in Deaths incidence HIV No. of adult from TB HIV/AIDS Incidence rate prevalence TB cases Adult TB Deaths in HIV- deaths New cases rate (per 2003­2004 in new attributable cases Deaths from TB infected TB deaths attributable Population of TB 100,000 (% per adult to HIV attributable from TB (per adults attributable to TB, 2003 (millions) (thousands) per year) year) cases (%) (thousands) of HIV (%) (thousands) 100,000) (thousands) to HIV (%) (%) WHO 722 2,573 356 3 33 600 28 587 81 206 32 11 African Region Rest of 5,665 6,345 112 1 3 141 3 1,106 20 43 4 8 the world Global 6,387 8,918 140 1 13 741 12 1,693 27 248 14 11 Source: WHO estimates for 2004, updated from Corbett et al. 2003. on finding smear-positive cases under DOTS and the new The increase in HIV prevalence has also been accompa- WHO Stop TB Strategy (Raviglione and Uplekar, forthcom- nied by a rise in the TB case-fatality rate, and hence the TB ing), the proportion of cases reported to be smear-positive death rate in the general population. One recent estimate has fallen in recent years in several African countries with put the fraction of AIDS deaths due to TB at 12 percent in high rates of HIV. Although there are uncertainties about the WHO African region in 2000 (Corbett et al. 2003), diagnosis, these data conform with the expectation that although this fraction could be higher. In an autopsy study there will be more smear-negative TB where there is more in Abidjan, Côte d'Ivoire, TB was found to be the cause HIV. Because HIV infection rates are higher in women than of death of 54 percent of patients with HIV infection or men, more TB cases are also being reported among women, AIDS (Lucas et al. 1993). Malawi has reported high early especially those between the ages of 15 and 24 years. TB case death rates of HIV-infected TB patients during the first reports are typically male-biased (WHO 2006), but in one to two months of treatment (Harries, Hargreaves, Gausi several African countries with high rates of HIV infection, et al. 2001). Whether this reflects late presentation and con- the majority of notified TB cases are now women. sequently severe TB disease or severe HIV-related illness, 184 | Christopher Dye, Anthony D. Harries, Dermot Maher, S. Mehran Hosseini, Wilfred Nkhoma, and Felix M. Salaniponi Figure 13.5 Prevalence of HIV in TB Cases (All Forms) in Surveys coordinated by the WHO and the International Relation to HIV Prevalence in Adults Age 15 to 49 Union against TB and Lung Disease (IUATLD) between 0.8 1996 and 2002 yielded data on anti-TB drug resistance among new and previously treated cases from sites in 10 countries in Sub-Saharan Africa (Espinal et al. 2001; WHO 0.6 cases 2004a). This limited number of surveys suggests that TB in MDR-TB is not a widespread problem in the region. Low 0.4 resistance rates (MDR-TB prevalence typically less than prevalence 3 percent among patients suffering a first episode of TB) 0.2 HIV could be explained by the recent introduction of rifampicin in Africa, by the use of rifampicin-free treatment regimens 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 in the continuation phase (during months three to eight), by HIV prevalence the growing use of directly observed treatment as recom- Source: Corbett et al. 2003. mended under the directly observed treatment, short course Note: Each point represents a different African population in which HIV prevalence was measured (DOTS) strategy, and by the use of fixed-dose combination both in TB patients and in the adult population. The fitted line estimates the incidence-rate ratio to be 5.8 (95 percent CL, 5.2­6.5), that is, the ratio of TB incidence rates in HIV-infected and HIV- tablets in a few countries (Espinal et al. 2001). uninfected populations. Implementation of the DOTS Strategy such as bacteremia or cryptococcal meningitis, is not for Tuberculosis Control known. The precise cause of death in patients with HIV- related TB has been difficult to determine because there Population Coverage, Case Detection, and Treatment have been so few autopsy studies. Outcome. The key components of the WHO Stop TB Although the ratio of TB incidence rates in HIV-infected Strategy are listed in box 13.1. The new strategy builds on and HIV-uninfected individuals is expected to vary during the foundations laid by DOTS, and the main aim is still to the course of the HIV epidemic (as the average level of prevent illness, transmission, and death by curing active TB immunocompetence declines), recent studies have shown cases (Raviglione and Uplekar, forthcoming; WHO 2006). that this incidence-rate ratio takes an average value of about With respect to the implementation of DOTS, the primary six (figure 13.5) (WHO 2002). Knowing both the incidence- goals of national TB programs are to detect 70 percent of rate ratio and the HIV infection rate in the general pop- new smear-positive cases arising each year and to success- ulation, we can calculate the proportion of people newly fully treat 85 percent of these. The target year was set by diagnosed with TB who are infected with HIV. Estimates WHO to be 2005, but the achievements made by 2005 vary widely between countries, from less than 1 percent on will not be fully known until the end of 2006. With the some African islands (for example, Comoros, Mauritius) to correct application of antituberculosis drugs (short-course over 50 percent in some countries, including Botswana, chemotherapy), it is possible to cure over 90 percent of new Malawi, South Africa, Zambia, and Zimbabwe. Overall, smear-positive TB patients who are neither resistant to first- about one-third (34 percent) of all adults who had TB in line drugs nor infected with HIV. Before the spread of HIV, Sub-Saharan Africa were infected with HIV in 2004. countries that met these two targets could expect to see a decline in TB incidence rates of 5 to 10 percent per year or more (Dye et al. 1998). Resistance to Antituberculosis Drugs By the end of 2004, the core DOTS strategy was available Drug resistance, and eventually multidrug resistance (MDR- in principle to 84 percent of people living in the WHO TB; that is, resistance to at least isoniazid and rifampicin), is African region (WHO 2006). The estimated case detection expected to occur whenever patients fail a course of anti-TB rate by DOTS programs was 48 percent, somewhat below chemotherapy. An assessment of the number and distribu- the global average of 53 percent, and not increasing as tion of drug-resistant TB cases is important for planning TB quickly as the global average (figure 13.6). However, much control, because the treatment of resistant cases is more uncertainty surrounds this assessment of case detection. costly and more complex if second-line drugs are used, and First, for countries that have estimates of the true TB inci- failures and deaths are more frequent. dence rate based on population surveys, these are typically Tuberculosis | 185 Box 13.1 The WHO Stop TB Strategy Vision: A world free of TB Goal: To dramatically reduce the global burden of TB by 2015 in line with the Millennium Development Goals and the Stop TB Partnership targets Objectives: · Achieve universal access to high-quality diagnosis and patient-centered treatment · Reduce the human suffering and socioeconomic burden associated with TB · Protect poor and vulnerable populations from TB, TB/HIV, and multidrug-resistant TB · Support development of new tools and enable their timely and effective use Targets: · MDG 6, Target 8: Halt and begin to reverse the incidence of TB by 2015 · Targets linked to the MDGs and endorsed by the Stop TB Partnership: ­ By 2005: detect at least 70% of new sputum smear-positive TB cases and cure at least 85% of these cases ­ By 2015: reduce TB prevalence and deaths rates by 50% relative to 1990 ­ By 2050: eliminate TB as a public health problem (1 case per million population) Components of the strategy and implementation approaches 1. Pursue high-quality DOTS expansion and enhancement a. Political commitment with increased and sustained financing b. Case detection through quality-assured bacteriology c. Standardized treatment with supervision and patient support d. An effective drug supply and management system e. Monitoring and evaluation system, and impact measurement 2. Address TB/HIV, MDR-TB, and other challenges ­ Implement collaborative TB/HIV activities ­ Prevent and control multidrug-resistant TB ­ Address prisoners, refugees, and other high-risk groups, and special situations 3. Contribute to health system strengthening ­ Actively participate in efforts to improve systemwide policy, human resources, financing, management, service delivery, and information systems ­ Share innovations that strengthen systems, including the Practical Approach to Lung Health (PAL) ­ Adapt innovations from other fields 4. Engage all care providers ­ Public­Public and Public­Private Mix (PPM) approaches ­ International Standards for Tuberculosis Care (ISTC) 5. Empower people with TB, and communities ­ Advocacy, communication, and social mobilization ­ Community participation in TB care ­ Patients' Charter for Tuberculosis Care 6. Enable and promote research ­ Programme-based operational research ­ Research to develop new diagnostics, drugs, and vaccines Source: Raviglione and Uplekar 2006. 186 | Christopher Dye, Anthony D. Harries, Dermot Maher, S. Mehran Hosseini, Wilfred Nkhoma, and Felix M. Salaniponi Figure 13.6 Progress toward the Target of 70 Percent Case to HIV coinfection in some countries, the failure to record Detection in the WHO African Region, Compared with the treatment outcomes is evidently a problem of program Average Progress Worldwide management. It would be extremely useful to have compre- (percent) hensive and reliable data on TB deaths and their trends in 100 African populations, but no country in Sub-Saharan Africa, 90 rate except South Africa, has a national system for recording and 80 WHO target reporting deaths by cause. 70 detection 60 50 Treatment of TB Patients Infected with HIV. HIV- case 40 positive TB patients suffer significant HIV-related 30 -positive morbidity during the course of TB treatment. Adverse 20 reactions to anti-TB drugs are more frequent and lead to smear 10 treatment interruptions and fatalities (Raviglione et al. 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 1997). Several studies in Africa have reported an increase in year recurrent TB in HIV-positive patients (Korenromp et al. WHO African Region World 2003), and national programs with good registration sys- Source: Data from WHO 2006. tems routinely record high rates of recurrent disease. HIV-positive TB patients have a much higher mortality rate during and after anti-TB treatment than HIV-negative tuberculin surveys of the prevalence (and hence, risk) of patients (Mukadi, Maher, and Harries 2001). This may infection carried out before the emergence of HIV change with wider use of antiretroviral (ARV) therapy, but (Cauthen, Pio, and ten Dam 2002). Few countries have clinical studies have shown that, without ARV therapy, 20 to surveyed the prevalence of infection during the last decade, 30 percent of HIV-positive and smear-positive PTB patients the exceptions being Kenya (Odhiambo et al. 1999) and die before the end of treatment, and about 25 percent of Tanzania (Tanzanian Tuberculin Survey Collaboration those who survive die during the following 12 months. The 2001). There are no recent national surveys in Africa of the immune status of a patient is an important predictor of prevalence of active TB. For the many African countries that death: lower CD4 cell counts at the time of diagnosis in have no survey data at all, the estimate of case detection is HIV-positive and smear-positive patients are associated with little more than an expert guess, based on what is known higher mortality rates (Graham, Coulter, and Gilks 2001). (mostly qualitatively) about the method of surveillance. HIV-positive patients who present with smear-negative PTB Second, the apparent upward trend in case detection could have higher case-fatality rates during treatment than those be explained partly by improved case finding and partly by who present with smear-positive PTB, probably because the real rise in incidence due to HIV. In sum, the data they are typically more immunosuppressed. describing incidence rates and their trends, and, hence, case- HIV-infected patients given drug regimens with no detection rates, are poor for most African countries. rifampicin have higher case-fatality rates, and higher relapse The outcome of treatment in the African region is some- rates, than those given regimens with rifampicin what clearer. The treatment success rate for more than (Korenromp et al. 2003). Rifampicin-containing regimens 480,000 smear-positive patients enrolled under DOTS in improve survival, possibly because they act more strongly 2003 was 72 percent (WHO 2006). Treatment success under against M. tuberculosis and, through the broad spectrum DOTS in Africa was low in part because the death rate was antibiotic activity of rifampicin, they may prevent other 7 percent, higher than in any other region of the world. bacterial infections. More important was the large proportion of patients for Notwithstanding the problems of implementation, whom the outcome of treatment was not known: 20 percent DOTS has made a bigger contribution to the management of patients defaulted from treatment, were transferred to of TB in Africa than any other strategy (for example, bacil- other treatment centers without follow-up, or were simply lus Calmette-Guérin [BCG] vaccination). DOTS has been not evaluated. It is highly likely that death was the outcome enlarged as the new Stop TB Strategy, in part to confront the for some patients recorded as defaulters or transfers. complexities that HIV adds to TB epidemiology (Raviglione Although the high reported death rates might be attributable and Uplekar, forthcoming; box 13.1). Tuberculosis | 187 Additional Measures to Control TB in the HIV/AIDS Era and IPT can reduce the short-term risk of TB in this group by about 60 percent, although there is little improvement The Stop TB Strategy must be implemented in Africa by in survival (Quigley et al. 2001). IPT may also be valuable improving the population coverage of DOTS and by adding for HIV-infected individuals even without tuberculin other key elements, including intensified TB case finding, testing. TB preventive treatment, HIV testing and ARV therapy for IPT probably protects HIV-positive people from active TB patients, and various interventions against HIV (and TB by reducing the risk of progression from recent and therefore indirectly against TB) (De Cock and Chaisson latent infection. Where the transmission rate of M. tubercu- 1999; Maher, Floyd, and Raviglione 2002). The Global Plan losis is relatively high, repeated exposure to infection proba- to Stop TB 2006­2015 (Stop TB Partnership and WHO bly accounts for the limited duration of benefit of up to 2006) is a blueprint for the implementation of the Stop TB 2.5 years (Quigley et al. 2001) following completion of a six- Strategy in all regions of the world, including Africa. month course of IPT. Not surprisingly, the duration of pro- Implementation of this wider strategy should complement tection depends on the duration of preventive treatment efforts to improve the basic tools for TB control, such as a (Fitzgerald, Morse et al. 2000). more efficacious vaccine (http://www.aeras.org; Young and Although cheap, IPT is at present used mostly for the Dye forthcoming), more accurate diagnostic tests (http:// protection of individuals, rather than to prevent transmis- www.finddiagnostics.org.), and better drugs for prevention sion. This is because children rarely develop infectious TB, and treatment (http://www.tballiance.org). and because it is hard to administer IPT to healthy adults on a large scale. Because IPT requires consumption of the drug Intensified TB Case Finding. DOTS has traditionally daily for at least six months, a process that is difficult for relied on passive case detection. It is possible that more cases health services and patients alike, many people who could could be found by searching more actively among certain benefit from treatment drop out before completion. The population groups, although the evidence that active case proportion of HIV-infected people who do complete a finding can yield more cases at acceptable cost remains course of IPT is typically small. For IPT to be effective in weak. Groups that might be targeted in Africa for improved preventing a large number of TB cases associated with HIV, case finding include people with respiratory symptoms it will be necessary to find ways of minimizing the dropout attending general hospitals (outpatients and inpatients); rate and to expand the provision of voluntary counseling health service providers and health care workers in the pub- services for HIV-positive patients (Hawken and Muhindi lic, private, and nongovernmental organization (NGO) sec- 1999). tors (Harries, Maher, and Nunn 1997); people attending IPT can also prevent TB from recurring in patients who centers for HIV testing and voluntary counseling ( Aisu et al. have already suffered one episode. Studies by Perriens and 1995); prisoners (Coninx et al. 2000); and household con- colleagues (1995) in the Democratic Republic of Congo tacts of those with infectious TB, including patients and (formerly Zaire) and by Fitzgerald, Desvarieux, and col- contacts known to be HIV positive (Nunn et al. 1994). The leagues (2000) in Haiti showed a higher rate of recurrent screening of child contacts, often neglected, can be an TB in HIV-infected individuals than in non-HIV-infected important benefit to individual children, although, because individuals treated with a six-month regimen containing children with TB are usually not infective to others, it will rifampicin throughout (the regimen used in the study in not decrease transmission (Topley, Maher, and Nyong'onya the Democratic Republic of Congo had a four-drug initial Mbewe 1996). phase and that in Haiti had a three-drug initial phase). In both studies, posttreatment prophylaxis (isoniazid and TB Preventive Treatment. Individuals at high risk of rifampicin in the study in the Democratic Republic of developing TB can benefit from preventive treatment, usu- Congo and isoniazid in the study in Haiti) decreased the ally six months of isoniazid. Isoniazid preventive treatment number of TB recurrences in HIV-positive patients but did (IPT) is recommended for children who are household not prolong survival. Based on these successes, further stud- contacts of an infectious case of TB and who, after screen- ies are needed before posttreatment prophylaxis can be used ing, are found not to have active TB themselves (Harries more widely; they must confirm the benefits, establish opti- and Maher 2004). Up to 15 percent of tuberculin-positive, mum regimens (drugs and duration), and assess operational HIV-positive adults will develop TB each year (WHO 1999), feasibility. 188 | Christopher Dye, Anthony D. Harries, Dermot Maher, S. Mehran Hosseini, Wilfred Nkhoma, and Felix M. Salaniponi BCG Immunization. Most of the 75 percent of infants in Figure 13.7 Expected Reductions in the Number of TB Cases Africa who were vaccinated with BCG in 2003 (WHO 2004b) in Kenya over 10 Years will be protected against disseminated and severe TB (for 10,000 example, meningeal and miliary TB) for the first few years yr of their lives (WHO 1995). The efficacy of BCG against 10 8,000 severe forms of TB in children is 70 to 80 percent, but it over 6,000 takes about 3,400 inoculations to prevent one case of averted meningitis and 9,300 inoculations to prevent one case of 4,000 miliary TB (Bourdin Trunz, Fine, and Dye forthcoming). cases TB However, most people who are vaccinated as children in of 2,000 no. Africa will not be protected against pulmonary TB as adults, 0 because the vaccine is unlikely to protect for longer than TB TB cure TLTI all lifelong HAART 80% HAART reduction detection HIV TLTI all adherence 100% in HIV 15 years and, in many populations, has low efficacy against HIV adherence Incidence adult pulmonary disease. Even though BCG is not expected Source: Currie et al. 2003. to have any significant impact in reducing transmission and Note: TLTI = treatment of latent TB infection; HAART = highly active antiretroviral therapy. incidence, the WHO recommends vaccination for all Reductions ( 95 percent CL) based on increase in coverage of each intervention by 1 percent over present values (i.e., 50 percent case detection, 70 percent cure, zero otherwise) and on neonates in Africa, except those with symptoms of HIV dis- stabilization of HIV prevalence at the estimated value of 14 percent among adults age 15 to 49. ease or AIDS (WHO 1996). Figure 13.8 Proportional Reduction in the Incidence of TB Interventions against HIV. In pilot projects, controlled over 20 Years among HIV-Positives as a Function of Effective Coverage and the CD4 Count per Microliter at Which trials, and national programs in less-developed countries, all People Start ARV Therapy of the following interventions have been shown to be effec- tive in preventing HIV infection: increased condom use, 800 0.6­0.8 treatment of sexually transmitted infections, reduction in 700 0.4­0.6 the number of sexual partners, safe injections, and drugs to therapy 600 prevent mother-to-child transmission of HIV (Merson, ARV 500 of 0.2­0.4 Dayton, and O'Reilly 2000). If HIV control programs can 400 start encourage the use of, or provide greater access to, these at 300 0.0­0.2 interventions, we can expect concomitant reductions in the count 200 + burden of HIV-related TB. However, in a comparative mod- CD4 100 eling analysis of DOTS and various other strategies to con- 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 trol TB and HIV-related TB, incidence and death rates were Coverage x compliance far more sensitive to improvements in TB case detection and cure than to the introduction of TB preventive therapy and Source: Adapted from Williams and Dye 2003. Note: Shaded areas of the graph show proportional reductions in lifetime risk of TB. ARV therapy ARV therapy, even when rates of HIV infection are high is assumed to reduce the incidence of TB to the level immediately after seroconversion. Even with 90 percent effective coverage, we can expect no more than a 20 percent reduction in incidence if (figure 13.7) (Currie et al. 2003). patients begin therapy at 200 CD4+ cells per microliter (lower point, at base of arrow). Beginning Although ARV therapy can prevent TB by preserving or at 500 CD4+ cells per microliter would prevent 40 to 60 percent of TB cases (upper point, at tip of arrow). restoring immunity, early therapy plus high levels of cover- age and compliance will be needed to avert a significant diagnosis of TB could provide an important entry point for fraction of TB cases (figure 13.8). The reason is that TB the treatment of HIV/AIDS. emerges as an AIDS-related illness at a median CD4 cell Besides efficacy, we must also consider affordability and count of about 250 per micrometer (Williams and Dye value for money. DOTS was known to be a relatively low- 2003). Beginning ARV therapy at 200 per microliter in the cost and cost-effective strategy to improve health before absence of an AIDS-related illness (WHO 2003) means that the emergence of HIV/AIDS (Murray et al. 1991). Among a high proportion of HIV-infected people that are destined the diversity of interventions available under the Stop TB to develop TB will progress to active disease before they are Strategy, the detection and treatment of active TB cases is offered ARV therapy. Nonetheless, ARV therapy could still the most cost-effective approach to TB control in Africa greatly extend the lives of HIV-infected TB patients, and the (Currie et al. 2005). Tuberculosis | 189 Box 13.2 Main Constraints to Improving TB Control, as Identified by National Program Managers in Africa Most important constraints Other constraints · Insufficient number of qualified staff · Inadequate financing · Inadequate preparation for decentralization as part · Limited access to TB diagnosis and treatment of health system reform · Inadequate laboratory network · Failure to engage all health care providers in DOTS · HIV/AIDS implementation · Limited public awareness of TB · Insufficient political commitment · Deficiencies in drug supply Source: WHO (2006) and preceding annual reports in the same series. Cotrimoxazole Prophylaxis. Prophylaxis against common incidence rates, the future of TB in Africa will remain unpre- intercurrent infections (for example, bacterial causes of dictable as long as the direction of HIV epidemics is also pneumonia and diarrhea and their complications) is another unknown. Will the significant reduction in HIV infection way to decrease the morbidity and mortality rates of HIV- documented in Uganda (Parkhurst 2002) be replicated across infected TB patients. Two studies in Côte d'Ivoire have eastern and southern Africa? Or can we expect HIV infection shown a beneficial effect of cotrimoxazole. One found that rates in other countries to stabilize at close to peak levels? the drug reduced deaths of HIV-infected TB patients by As the spread of HIV infection and AIDS adds to the com- 48 percent; the other showed a significant reduction in mor- plexity of health care, studies that can identify the limiting bidity, but not mortality (Anglaret et al. 1999; Wiktor et al. factors in TB control will have great value. The national TB 1999). As a result, cotrimoxazole prophylaxis has been pro- program managers of some African countries have visionally recommended for HIV-infected individuals in attempted to identify the main constraints to improving Africa as part of the minimum package of care (UNAIDS their program's performance (WHO 2006), but the list in 2001). In Malawi, voluntary counseling and HIV testing box 13.2 needs to be refined, quantified, and customized for for TB patients, with treatment including cotrimoxazole, different countries. More money is needed for TB control, as reduced case-fatality rates by almost 20 percent compared set out in the Global Plan to Stop TB 2006­2015 (Stop TB with a control group (Zachariah et al. 2003). Further studies Partnership and WHO 2006) but money is not the only in other sites are necessary to confirm and evaluate the essential commodity. The mechanisms to identify other nec- benefits and the duration of effectiveness, and the feasibility essary materials and processes will include better analyses of of using cotrimoxazole under routine conditions. the abundant surveillance data that have already been col- lected by national TB programs. Good operational research, which need not be costly or complex, will also suggest ways CONCLUSION to improve the performance of the programs. For example, following investigations in several African countries, it is Before HIV infection and AIDS emerged and spread in now clear that community-based care can give, under a wide Africa, TB incidence rates were typically under 100 per range of circumstances, satisfactory treatment results at 100,000 persons per year and falling. HIV has turned a slow lower cost (Floyd et al. 2003; Moalosi et al. 2003; Nganda et decline into a rapid resurgence, especially in eastern and al. 2003; Okello et al. 2003; Sinanovic 2003). Pragmatic field southern Africa. However, the worst HIV epidemics are now studies must continue to explore better ways of using the almost certainly decelerating, and even turning downward in current tools for TB control as new vaccines, drugs, and some countries (UNAIDS and WHO 2005). Because the diagnostics begin to emerge from laboratories. interval between HIV infection and the onset of TB is four Whatever the impact of HIV on TB in the next few years, to six years, we can expect TB incidence rates to continue African countries will continue to need vigorous TB control increasing for some years in some African countries. But we programs that fully implement the new Stop TB Strategy, are now, at least, in a position to evaluate the maximum size founded on DOTS. Even with high rates of HIV infection, of the TB problem created by HIV in Africa. Beyond peak TB DOTS implementation is relatively cheap and cost-effective. 190 | Christopher Dye, Anthony D. Harries, Dermot Maher, S. Mehran Hosseini, Wilfred Nkhoma, and Felix M. Salaniponi Other methods for the prevention and treatment of HIV Di Perri, G., M. Cruciani, M. C. Danzi, R. Luzzati, G. De Checchi, M. Malena, S. Pizzighella, et al. 1989. "Nosocomial Epidemic of Active and AIDS will be needed too, but they should be introduced Tuberculosis in HIV Infected Patients." Lancet 2: 1502­4. in ways that will be complementary to DOTS. Tuberculosis Dye, C., G. P. Garnett, K. Sleeman, and B. G. Williams. 1998. "Prospects for and HIV control programs now clearly have mutual con- Worldwide Tuberculosis Control under the WHO DOTS Strategy. cerns: the prevention of HIV infection and the treatment of Directly Observed Short-Course Therapy." Lancet 352: 1886­91. Dye, C., S. Scheele, P. Dolin, V. Pathania, and M. C. Raviglione. 1999. AIDS should be components of TB control, and TB care "Global Burden of Tuberculosis: Estimated Incidence, Prevalence and and prevention should be priorities in the management of Mortality by Country." Journal of the American Medical Association 282: HIV/AIDS. Until recently, TB programs and HIV/AIDS pro- 677­86. grams have pursued separate courses. They can no longer Espinal, M. A., A. Laszlo, L. Simonsen, F. Boulahbal, S. J. Kim, A. Reniero, S. Hoffner, et al. 2001. "Global Trends in Resistance to Anti-TB Drugs." afford to do so. New England Journal of Medicine 344: 1294­1303. Ewer, K., J. Deeks, L. Alvarez, G. Bryant, S. Waller, P. Andersen, P. Monk, and A. Lalvani. 2003. "Comparison of T-Cell-Based Assay with Tuberculin Skin Test for Diagnosis of Mycobacterium Tuberculosis REFERENCES Infection in a School Tuberculosis Outbreak." Lancet 361: 1168­73. Fitzgerald, D. W., M. Desvarieux, P. Severe, P. Joseph, W. D. Johnson Jr., and Ackah, A. N., D. Coulibaly, H. Digbeu, K. Diallo, K. M.Vetter, I. M. Coulibaly, J. W. Pape. 2000. "Effect of Post-Treatment Isoniazid on Prevention of A. E. Greenberg, and K. M. De Cock. 1995. "Response to Treatment, Recurrent Tuberculosis in HIV-1-Infected Individuals: A Randomised Mortality, and CD4 Lymphocyte Counts in HIV-Infected Persons with Trial." Lancet 356: 1470­74. Tuberculosis in Abidjan, Cote d'Ivoire." Lancet 345: 607­10. Fitzgerald, D. W., M. M. Morse, J. W. Pape, and W. D. Johnson Jr. 2000. Aisu, T., M. C. Raviglione, E. van Praag, P. Eriki, J. P. Narain, L. Barugahare, "Active Tuberculosis in Individuals Infected with Human G. Tembo, D. McFarland, and F. A. Engwau. 1995. "Preventive Immunodeficiency Virus after Isoniazid Prophylaxis." Clinical Chemotherapy for HIV-Associated Tuberculosis in Uganda: An Infectious Diseases 31: 1495­97. Operational Assessment at a Voluntary Counselling and Testing Floyd, K., J. Skeva, T. Nyirenda, F. Gausi, and F. Salaniponi. 2003. "Cost and Centre." AIDS 9: 267­73. Cost-Effectiveness of Increased Community and Primary Care Facility Anglaret, X., G. Chene, A. Attia, S. Toure, S. Lafont, P. Combe, K. Manlan, Involvement in Tuberculosis Care in Lilongwe District, Malawi." T. N'Dri-Yoman, and R. Salamon. 1999."Early Chemoprophylaxis with International Journal of Tuberculosis and Lung Disease 7 (9 Suppl. 1): Trimethoprim-Sulphamethoxole for HIV-1-Infected Adults in S29­37. Abidjan, Côte d'Ivoire: A Randomized Trial." Lancet 353: 1463­68. Graham, S. M., J. B. S. Coulter, and C. F. Gilks. 2001. "Pulmonary Disease Bourdin Trunz, B., P. E. M. Fine, and C. Dye. Forthcoming. "Global Impact in HIV-Infected African Children." International Journal of Tuberculosis of BCG Vaccination on Childhood Tuberculous Meningitis and and Lung Disease 5: 12­23. Miliary Tuberculosis." Lancet. Hargreaves, N. J., O. Kadzakumanja, S. Phiri, C. H. Lee, X. Tang, F. M. Cauthen, G. M., A. Pio, and H. G. ten Dam. 2002. "Annual Risk of Salaniponi, A. D. Harries, and S. B. Squire. 2001. "What Causes Smear- Tuberculous Infection. 1988." Bulletin of the World Health Organization Negative Pulmonary Tuberculosis in Malawi, an Area of High HIV 80: 503­11. Seroprevalence?" International Journal of Tuberculosis and Lung Disease Comstock, G. W., V. T. Livesay, and S. F. Woolpert. 1974. "The Prognosis of 5: 1­10. a Positive Tuberculin Reaction in Childhood and Adolescence." Harries, A. D., N. J. Hargreaves, F. Gausi, J. H. Kwanjana, and F. M. American Journal of Epidemiology 99: 131­38. Salaniponi. 2001. "High Early Death Rate in Tuberculosis Patients Coninx, R., D. Maher, H. Reyes, and M. Grzemska. 2000. "Tuberculosis in in Malawi." International Journal of Tuberculosis and Lung Disease 5: Prisons in Countries with High Prevalence." British Medical Journal 1000­05. 320: 440­42. Harries, A. D., N. J. Hargreaves, S. M. Graham, C. Mwansambo, Corbett, E. L., C. J. Watt, N. Walker, D. Maher, B. G. Williams, M. C. P. Kazembe, R. L. Broadhead, D. Maher, and F. M. Salaniponi. 2001. Raviglione, and C. Dye. 2003. "The Growing Burden of Tuberculosis: "Childhood Tuberculosis in Malawi: Nationwide Case-Finding and Global Trends and Interactions with the HIV Epidemic." Archives of Treatment Outcomes." International Journal of Tuberculosis and Lung Internal Medicine 163: 1009­21. Disease 6: 424­31. Crofton, J., N. Horne, and F. Miller. 1999. Clinical Tuberculosis, 2nd ed. Harries, A. D., and D. Maher. 2004. TB/HIV: A Clinical Manual. 2nd ed. London: Macmillan Education. WHO/HTM/TB/2004.329. Geneva: WHO. Currie, C. S., K. Floyd, B. G. Williams, and C. Dye. 2005. "Cost, Harries, A. D., D. Maher, and P. Nunn. 1997. "Practical and Affordable Affordability and Cost-effectiveness of Strategies to Control Measures for the Protection of Health Care Workers from Tuberculosis Tuberculosis in Countries with High HIV Prevalence." BMC Public in Low-Income Countries." Bulletin of the World Health Organization Health 5: 130. 75: 477­89. Currie, C. S. M., B. G. Williams, R. C. Cheng, and C. Dye. 2003. Harries, A. D., L. Nyong'Onya Mbewe, F. M. Salaniponi, D. S. Nyangulu, "Tuberculosis Epidemics Driven by HIV: Is Prevention Better Than J. Veen, T. Ringdal, and P. Nunn. 1996. "Tuberculosis Programme Cure?" AIDS 17: 2501­8. Changes and Treatment Outcomes in Patients with Smear-Positive Daley, C. L. 1993. "Tuberculosis Recurrence in Africa: True Relapse or Pulmonary Tuberculosis in Blantyre, Malawi." Lancet 347: 807­9. Re-infection?" Lancet 342: 756­57. Hawken, M., and D. W. Muhindi. 1999. "Tuberculosis Preventive Therapy De Cock, K. M., and R. E. Chaisson. 1999. "Will DOTS Do It? A in HIV-Infected Persons: Feasibility Issues in Developing Countries." Reappraisal of Tuberculosis Control in Countries with High Rates of International Journal of Tuberculosis and Lung Disease 3: 646­50. HIV Infection." International Journal of Tuberculosis and Lung Disease Kenyon, T. A., M. J. Mwasekaga, R. Huebner, D. Rumisha, N. Binkin, and 3: 457­65. E. Maganu. 1999. "Low Levels of Drug-Resistance Amidst Rapidly Tuberculosis | 191 Increasing Tuberculosis and Human Immunodeficiency Virus Patients in Rural Uganda." International Journal of Tuberculosis and Co-Epidemics in Botswana." International Journal of Tuberculosis and Lung Disease 7 (9 Suppl. 1): S72­79. Lung Disease 3: 4­11. Parkhurst, J. O. 2002. "The Ugandan Success Story? Evidence and Claims Korenromp, E. L., F. Scano, B. G. Williams, C. Dye, and P. Nunn. 2003. of HIV-1 Prevention." Lancet 360: 78­80. "Effects of Human Immunodeficiency Virus Infection on Recurrence Perriens, J. H., M. E. St Louis, Y. B. Mukadi, C. Brown, J. Prignot, of Tuberculosis after Rifampicin-Based Treatment: An Analytical F. Pouthier, F. Portaels, et al. 1995. "Pulmonary Tuberculosis in HIV- Review." Clinical Infectious Diseases 37: 101­12. Infected Patients in Zaire. A Controlled Trial of Treatment for Either 6 Lienhardt, C., and L. C. Rodrigues. 1997. "Estimation of the Impact of the or 12 Months." New England Journal of Medicine 332: 779­84. Human Immunodeficiency Virus Infection on Tuberculosis: Quigley, M. A., A. Mwinga, M. Hosp, I. Lisse, D. Fuchs, J. D. H. Porter, and Tuberculosis Risks Revisited?" International Journal of Tuberculosis and P. Godfrey-Faussett. 2001. "Long Term Effect of Preventive Therapy for Lung Disease 1: 196­204. Tuberculosis in a Cohort of HIV-Infected Zambian Patients." AIDS 15: Lucas, S. B., K. M. De Cock, A. Hounnou, C. Peacock, M. Diomande, 215­22. M. Honde, A. Beaumel, L. Kestens, and A. Kadio. 1994. "The Raviglione, M. C., A. D. Harries, R. Msiska, D. Wilkinson, and P. Nunn. Contribution of Tuberculosis to Slim Disease in Africa." British Medical 1997. "Tuberculosis and HIV: Current Status in Africa." AIDS 11 Journal 308: 1531­33. (Suppl. B): S115­23. Lucas, S. B., A. Hounnou, C. Peacock, A. Beaumel, G. Djomand, J. M. Raviglione, M. C., and M. W. Uplekar. Forthcoming. "The New Stop TB N'Gbichi, K. Yeboue, et al. 1993. "The Mortality and Pathology of HIV Strategy of WHO." Lancet. Infection in a West African City." AIDS 7: 1569­79. Maher, D., K. Floyd, and M. Raviglione. 2002. "Strategic Framework to Richter, C., B. Ndosi, A. S. Mwammy, and R. K. Mbwambo. 1991. Decrease the Burden of TB/HIV." WHO/CDS/TB/2002.296, WHO, "Extrapulmonary Tuberculosis--a Simple Diagnosis?" Tropical and Geneva. Geographical Medicine 43: 375­78. Mazurek, G. H., and M. E. Villarino. 2003. "Guidelines for Using the Rieder, H. L. 1999. Tuberculosis Epidemiology. 1st ed. Paris: IUATLD. QuantiFERON-TB Test for Diagnosing Latent Mycobacterium Rieder, H. L., G. M. Cauthen, G. W. Comstock, and D. E. Snider. 1989. Tuberculosis Infection." Morbidity and Mortality Weekly Report 52 "Epidemiology of Tuberculosis in the United States." Epidemiologic (RR-2): 15­18. Reviews 11: 79­98. Merson, M. H., J. M. Dayton, and K. O'Reilly. 2000. "Effectiveness of Sinanovic, E., K. Floyd, L. Dudley, V. Azevedo, R. Grant, and D. Maher. HIV Prevention Interventions in Developing Countries." AIDS 14 2003. "Cost and Cost-Effectiveness of Community-Based Care for (Suppl. 2), S68­84. Tuberculosis in Cape Town, South Africa." International Journal of Moalosi, G., K. Floyd, J. Phatshwane, T. Moeti, N. Binkin, and T. Kenyon. Tuberculosis and Lung Disease 7 (9 Suppl. 1): S56­62. 2003. "Cost-Effectiveness of Home-Based Care versus Hospital Care Sonnenberg, P., J. Murray, J. R. Glynn, S. Shearer, B. Kambashi, and for Chronically Ill Tuberculosis Patients, Francistown, Botswana." P. Godfrey-Faussett. 2001. "HIV-1 and Recurrence, Relapse, and International Journal of Tuberculosis and Lung Disease 7 (9 Suppl. 1): Reinfection of Tuberculosis after Cure: A Cohort Study in South S80­85. African Mineworkers." Lancet 358: 1687­93. Mukadi, Y. D., D. Maher, and A. Harries. 2001. "Tuberculosis Case Fatality Stop TB Partnership and WHO (World Health Organization). 2006. The Rates in High HIV Prevalence Populations in Sub-Saharan Africa." Global Plan to Stop TB, 2006­2015. WHO/HTM/STB/2006.35. Geneva: AIDS 15: 143­52. WHO. Mukadi, Y., J. H. Perriens, M. E. St. Louis, C. Brown, J. Prignot, J. C. Styblo, K. 1991. Epidemiology of Tuberculosis. Selected Papers. Vol. 24. The Willame, F. Pouthier, et al. 1993. "Spectrum of Immunodeficiency in Hague: Royal Netherlands Tuberculosis Association KNCV. HIV-1-Infected Patients with Pulmonary Tuberculosis in Zaire." Lancet 342: 143­46. Sutherland, I. 1976. "Recent Studies in the Epidemiology of Tuberculosis, Based on the Risk of Being Infected with Tubercle Bacilli." Advances in Murray, C. J., E. DeJonghe, H. J. Chum, D. S. Nyangulu, A. Salomao, and Tuberculosis Research 19: 1­63. K. Styblo. 1991. "Cost Effectiveness of Chemotherapy for Pulmonary Tuberculosis in Three Sub-Saharan African Countries." Lancet 338: Tanzania Tuberculin Survey Collaboration. 2001. "Tuberculosis Control 1305­8. in the Era of the HIV Epidemic: Risk of Tuberculosis Infection in Tanzania, 1983­98." International Journal of Tuberculosis and Lung Nganda, B., J. Wang'ombe, K. Floyd, and J. Kangangi. 2003. "Cost and Disease 5: 103­12. Cost-Effectiveness of Increased Community and Primary Care Facility Involvement in Tuberculosis Care in Machakos District, Kenya." Topley, J., D. Maher, and L. Nyong'onya Mbewe. 1996. "Transmission of International Journal of Tuberculosis and Lung Disease 7 (9 Suppl. 1): Tuberculosis to Contacts of Sputum Positive Adults in Malawi." S14­20. Archives of Disease in Childhood 74: 140­43. Nunn, P., A. M. Elliott, and K. P. W. J. McAdam. 1994. "Impact of Human UNAIDS (Joint United Nations Programme on HIV/AIDS). 2001. Immunodeficiency Virus on Tuberculosis in Developing Countries." "Provisional Recommendations on the Use of Cotrimoxazole as Part of Thorax 49: 511­18. a Minimum Package of Care in Adults and Children Living with HIV/AIDS in Africa." Nunn, P., M. Mungai, J. Nyamwaya, C. Gicheha, R. J. Brindle, D. T. Dunn, W. Githui, J. O. Were, and K. P. McAdam. 1994. "The Effect of Human UNAIDS (Joint United Nations Programme on HIV/AIDS) and WHO Immunodeficiency Virus Type 1 on the Infectiousness of Tuberculosis." (World Health Organization). 2005. AIDS Epidemic Update, December Tubercle and Lung Disease 75: 25­32. 2005. UNAIDS/05.19E. Geneva: UNAIDS and WHO. Odhiambo, J. A., M. W. Borgdorff, F. M. Kiambih, D. K. Kibuga, D. O. Von Pirquet, C. 1909."Frequency of Tuberculosis in Childhood." Journal of Kwamanga, L. Ng'ang'a, R. Agwanda, et al. 1999. "Tuberculosis and the the American Medical Association 52: 675­78. HIV Epidemic: Increasing Annual Risk of Tuberculous Infection in WHO (World Health Organization). 1995. "Global Tuberculosis Kenya, 1986­1996." American Journal of Public Health 89: 1078­82. Programme and Global Programme on Vaccines." Statement on Okello, D., K. Floyd, F. Adatu, R. Odeke, and G. Gargioni. 2003. "Cost and BCG Revaccination for the Prevention of Tuberculosis. Weekly Cost-Effectiveness of Community-Based Care for Tuberculosis Epidemiological Record 70: 229­36. 192 | Christopher Dye, Anthony D. Harries, Dermot Maher, S. Mehran Hosseini, Wilfred Nkhoma, and Felix M. Salaniponi ------. 1996. Immunisation Policy. Global Programme on Vaccines. ------. 2006. Global Tuberculosis Control: Surveillance, Planning and Geneva: WHO. Financing. WHO/HTM/TB/2006.362. Geneva: WHO. ------. 1999. "Preventive Therapy against Tuberculosis in People Living Wiktor, S. Z., M. Sassan-Morokro, A. D. Grant, L. Abouya, J. M. Karon, C. with HIV." Weekly Epidemiological Record 74: 385­98. Maurice, G. Djomand, et al. 1999. "Efficacy of Trimethoprim- Sulphamethoxazole Prophylaxis to Decrease Morbidity and Mortality ------. 2000. The Economic Impacts of Tuberculosis. Stop TB Initiative in HIV-1-Infected Patients with Tuberculosis in Abidjan, Côte d'Ivoire: 2000 Series. WHO/CDS/STB/2000.5. Geneva: WHO. A Randomised Controlled Study." Lancet 353: 1469­75. ------. 2002. The World Health Report--Reducing Risks, Promoting Williams, B. G., and C. Dye. 2003. "Antiretroviral Drugs for Tuberculosis Healthy Life. Geneva: WHO. Control in the Era of HIV/AIDS." Science 301: 1535­37. ------. 2003. "Scaling Up Anti-Retroviral Therapy in Resource-Limited Young, D. B., and C. Dye. Forthcoming. "The Development and Impact of Settings: Treatment Guidelines for a Public Health Approach." WHO, Tuberculosis Vaccines." Cell. Geneva. Zachariah, R., M-P. Spielmann, C. Chinji, P. Gomani, V. Arendt, N. J. ------. 2004a. Anti-Tuberculosis Drug Resistance in the World. Report Hargreaves, F. M. L. Salaniponi, and A. D. Harries. 2003. "Voluntary Number 3. WHO/HTM/TB/2004.343. Geneva: WHO. Counselling, HIV Testing and Adjunctive Cotrimoxazole Reduces ------. 2004b. "WHO Vaccine-Preventable Disease Monitoring System. Mortality in Tuberculosis Patients in Thyolo, Malawi." AIDS 17: 2004 Global Summary." WHO/IVB/2004, WHO, Geneva. 1053­61. Tuberculosis | 193 Chapter 14 Malaria Robert W. Snow and Judy A. Omumbo Plasmodium falciparum is the most common of the four losis, infection with the malaria parasite is almost always human malaria parasites across much of Sub-Saharan universal in a population, and the presence of the pathogen Africa. (The other three parasites are P. vivax, P. malariae, is not a sufficient marker of disease. Individuals who die and P. ovale.) The distribution of P. vivax is concentrated in from malaria represent the public health costs of develop- the Horn of Africa, covering Djibouti, Eritrea, Ethiopia, ing immunity at a population level. These deaths are Somalia, and Sudan. P. falciparum accounts for almost all concentrated among those with poorly developed immuni- the malaria mortality in Sub-Saharan Africa, and it is often ty, and, generally, young children bear the brunt of the stated that the continent bears over 90 percent of the global mortality burden. Individuals born into areas of stable P. falciparum burden. Recent bioinformatics analysis of P. falciparum transmission frequently acquire and clear changes in human ecology suggest that about 6,000 years infections without becoming ill, but most will, at some ago, P. falciparum populations expanded rapidly in Africa stage in their lives, develop an overt clinical response to and spread worldwide, coincident with human population infection, often manifested as fever. These clinical events growth and subsequent diasporas facilitated by the dawn of may lead to severe complications, which may resolve natu- agriculture (Joy et al. 2003). This parasite has exacted a rally, require medical intervention, or result in death. heavy mortality toll on Africa's population, evidenced by the Figure 14.1 shows a typical age-structured series of risks of selection for several human survival mechanisms, such as infection, mild clinical disease, severe disease, and death the genetic polymorphisms associated with red cell struc- due to P. falciparum for a population living on the Kenyan ture and function (Hill 1992). coast (Snow and Gilles 2002). Many individuals naturally Malaria infection is common in Sub-Saharan Africa, but acquire functional immune responses to severe disease and death directly attributed to the parasite is comparatively death early in life; immunity to the milder consequences rare, largely because of acquired functional immunity. of infection occurs later in childhood, but the ability to Unlike the human immunodeficiency virus (HIV) and the sterilize blood-stage infection probably does not occur acquired immune deficiency syndrome (AIDS) or tubercu- until adulthood. 195 Figure 14.1 Malarial Risks in Children Age 0 to 14 Years in a unlikely to affect health outcomes. Understanding this Stable Endemic Area of the Kenyan Coast relationship is important for defining the age-specific mortality burdens in Sub-Saharan Africa, an area able to support infection rates ranging from one infection every risk three years to hundreds of new infections per year (Hay et al. 2000). relative In addition to the morbidity and mortality directly attributed to P. falciparum, other consequential and indirect 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 effects are linked to each step of the infection and disease age process. Chronic, subclinical infections cause anemia or Relative risk of infection Risk of morbidity may encourage undernutrition, which in turn may increase Risk of severe disease Risk of mortality susceptibility to severe clinical outcomes of subsequent malarial or other pathogenetic infection. During pregnancy, Source: Snow and Gilles 2002. Note: Scale of y axis has arbitrarily been fitted to demonstrate relative change in risk by age. asymptomatic infection of the placenta significantly reduces birthweights and infant survival rates. Patients who survive The relation between the frequency of parasite exposure severe disease may be left with debilitating sequelae, such as and disease outcome is complex. The speed with which a spasticity or epilepsy. Subtler consequences include behav- population acquires functional immunity to the severe con- ioral disturbances or cognitive impairment (Holding and sequences of P. falciparum infection depends on the fre- Snow 2001; Holding and Wekulo 2004). These combined quency of parasite exposure from birth as measured by the effects are summarized in figure 14.2. In the absence of intensity of parasite transmission in a given locality (see measures aimed at reducing the risk of infection, the risks Snow and Marsh 2002). The shape of the severe disease and shown in figure 14.2 largely depend on extrinsic factors, mortality curves shown in figure 14.1 may therefore be such as those that determine the speed with which a popu- shifted to the right for areas where parasite transmission is lation develops acquired immunity and the access to effec- of low, stable intensity and to the left for areas where trans- tive case management, and on intrinsic factors, such as host mission is of high, stable intensity. Where infection is rare genetics. the risk of mortality is likely to be directly related to the risk This chapter describes the determinants and distribution of infection, because acquired functional immunity is of P. falciparum infection risk in Sub-Saharan Africa and Figure 14.2 Public Health Effects of Plasmodium falciparum Malaria Treatment side effects · Inappropriate management by unqualified practitioners · Adverse drug reactions of antimalarials or antipyretics · HIV risks from blood transfusions Cognitive impairment/ behavioral disturbance Sequelae Clinical Severe clinical Uninfected Infected disease complications Blindness Hearing or speech impairments Death Spasticity Epilepsy Anemia Undernutrition Low birthweight Increased susceptibility to general infection Source: Snow and Gilles 2002. 196 | Robert W. Snow and Judy A. Omumbo uses populations at risk to estimate mortality from malaria. Rainfall is also related to humidity and saturation deficit, It also considers the evidence for consequential and indirect both affecting mosquito survival (adult vector longevity mortality and describes P. falciparum as a risk factor for increases with humidities over 60 percent). rather than a cause of pediatric mortality. The chapter concludes with a description of the relationship between poverty and malaria and recent trends in malaria mortality Mapping Malaria Risks in Sub-Saharan Africa to provide some context for current Using the climatic determinants of transmission identified international efforts to halve the malaria burden by the year above, the Mapping Malaria in Africa Project (http:// 2010 (Nabarro and Taylor 1998). www.mara.co.za) developed a series of risk maps for stable P. falciparum transmission across the continent (Craig, Snow, and le Sueur 1999). In brief, the project used long- PLASMODIUM FALCIPARUM DISTRIBUTION term mean monthly temperature and rainfall data to define IN AFRICA the limits of distribution of stable endemic malaria across Climate, local ecology, and active control affect the ability of Africa. Temperature and rainfall profiles in sample areas, malaria parasites and their anopheline mosquito vectors to where malaria endemicity was known, were translated into coexist long enough to enable transmission. The frequency a model of "climate suitability." The temperature limits were of transmission, or endemicity, depends on the density and related to the requirements for the extrinsic parasite devel- infectivity of anopheline vectors. These features depend on opment cycle. The model made the following assumptions: a range of climatic, physical, and population characteristics, (a) the optimal temperature range was 18°C to 22°C; for example, rainfall, location of human settlements near or (b) optimal rainfall values were greater than or equal to at rivers or other mosquito larval breeding sites, and the 80 millimeters; (c) conditions of rainfall and temperature density of human populations in a village. The most signif- had to coincide on a month-to-month basis for at least five icant determinant of the intensity of parasite transmission is consecutive months (three months for the northern fringes climate. of Sub-Saharan Africa); and finally, (d) a frost factor (mean monthly minimum temperature of less than 5°C for any one month) was used to eliminate transmission at any point. Climate Determinants of P. falciparum Transmission in Africa The model provided fuzzy climate suitability (FCS) values, The development of both the vector and parasite is temper- ranging from zero (unsuitable, hence malaria absent) to one ature dependent. The optimum temperature range for par- (very suitable, malaria endemic). asite development in the female Anopheles (sporogony) is During earlier attempts to describe the malaria burden in between 25°C and 30°C, and development ceases below Sub-Saharan Africa the climate suitability maps for P. falci- 16°C. Intermittent low temperatures delay sporogony, and parum transmission were combined with interpolated maps the period immediately after the infective bite by the mos- of population distribution (Deichmann 1996; Snow et al. quito on an infected human host is the most sensitive to 1999). The population data were initially constructed using drops in temperature. Above 35°C sporogony slows down population totals from the last available censuses for admin- considerably. Extremely high temperatures are associated istrative units (communities, towns, or districts). These data with the development of smaller and less fecund adult mos- were then converted into a regular raster grid of population quitoes. Thermal death of mosquitoes occurs at 40°C to totals, and auxiliary information was used to distribute the 42°C. Altitude and temperature are strongly correlated: with population within the administrative unit across its raster every 100-meter increase in altitude, the temperature drops grid cells. The process heuristically incorporated informa- by 0.5°C. Overall, the use of altitude as a marker of endemic- tion on where people tend to live: in or close to towns and ity or disease risk is vague, yet there is a tendency within the cities, close to transportation infrastructure, around pro- literature to refer to highland malaria in East Africa and the tected areas, near water bodies, and not at very high eleva- Horn of Africa. tions. Using the Geographical Information System (GIS)­ Numerous studies have demonstrated the association based information on the location and size of towns and between Anopheles gambiae sensus lato (the most important cities, roads, railroads, navigable rivers, and uninhabitable vector of P. falciparum in Africa) abundance and rainfall. areas, population density was weighted, a high value imply- Without surface water the female Anopheles cannot lay eggs. ing a high density and a low or zero value implying low Malaria | 197 or no population. These weights were then used to Figure 14.3 Fuzzy Climate Suitability Membership for Malaria proportionately distribute population to grid cells. The dig- 30° 20° 10° 0° 10° 20° 30° 40° 50° IBRD 34533 MARCH 2006 ital map extracted population distributions according to Mediterranean Sea each cell in a regular raster grid with a resolution of 5 kilo- 30° 30° meters at the equator. It was then combined with the maps Red of malaria risk to provide simple population totals of those 20° Sea 20° exposed to stable, unstable, or no malaria risk. The P. falciparum risk-to-population distribution for 10° 10° Africa has subsequently been refined through use of an improved link between mortality data and transmission 0° ATLANTIC 0° intensity (Snow et al. 2003). New criteria were adopted fol- OCEAN INDIAN OCEAN lowing an improved understanding of the variations in dis- 10° 10° ease outcomes and risks congruent with variations in stable P. falciparum transmission. The relation between the exact 20° 20° number of new infections a population is exposed to each no transmission or population (FCS = 0) year (annual entomological inoculation rates, or EIR) and negligible malaria risk (FCS > 0 and < 0.25) low stable endemic risk or epidemic prone parasite prevalence is nonlinear, but quartiles of the preva- 30° (FCS > 0.25 and < 0.75) 30° stable endemic malaria areas (FCS > 0.75) lence of infection have been shown to approximate loga- no data international boundaries This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank rithmic increases in the EIR (Beier, Killeen, and Githure Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° 1999). This principle has been used previously to categorize Source: Modified from Craig, Snow, and Ie Sueur 1999. pediatric malaria and all-cause mortality according to para- Note: Areas of no transmission or population: class1, FCS 0; areas of negligible malaria risk: class 2, FCS 0 and 0.25; areas of low, stable endemic risk or epidemic prone: class 3, site prevalence estimates from childhood cross-sectional FCS 0.25 and 0.75; and stable endemic malaria areas: class 4, FCS 0.75. surveys (Snow, Korenromp, and Gouws 2004; Snow and Marsh, 2002). When infection prevalence quartiles (less than 25 percent, 25 to 49 percent, 50 to 74 percent, and are represented by a FCS greater than zero but less than 0.25 75 percent and greater) are used, mortality indicators satu- (class 2). Parasite prevalence rates are likely to be extremely rate at the highest two classes of endemicity and rise sharply low or zero; however, these populations often border areas from the x­y intercept within the first class. Analysis of 217 of higher infection risk and may be subject to aberrations in independent parasitological surveys among children in climate that lead to localized transmission. Third, in some Kenya suggests that most areas (81 percent) with a FCS areas populations are exposed to a FCS value between 0.25 value of less than 0.75 are represented by communities sup- and 0.75 (class 3). Communities living in these areas are porting parasite prevalence rates of less than 25 percent likely to support cross-sectional parasite prevalence esti- (Omumbo et al. 2004). It seems appropriate, therefore, to mates in childhood of less than 25 percent, and transmission distribute the population of Sub-Saharan Africa into areas is likely to be acutely seasonal with a tendency toward epi- of no risk, unstable transmission risk, low stable risk, and demics. In practice it would be hard to distinguish between moderate to high stable endemic risk. classes 2 and 3 in many areas of Sub-Saharan Africa outside The fuzzy membership model developed by Craig, Snow, of southern Africa, and therefore they are grouped together and le Sueur (1999) was used to delineate four ecologically here. Finally, some areas of the continent have stable endem- distinct areas of Africa (figure 14.3). First, there are areas ic malaria transmission (FCS greater than or equal to 0.75), with no human settlement, or FCS values of zero (class 1). which may still vary seasonally, but support parasite preva- Some countries have minimal and only very localized risks lence rates greater than or equal to 25 percent (class 4). In of malaria transmission, with over 90 percent of their pop- southern Africa (Botswana, Namibia, South Africa, ulation residing in class 1 areas, the western Sahara region Swaziland, and Zimbabwe) a different approach was taken, (Algeria, the Arab Republic of Egypt, Lesotho, Libya, because classes 2 and 3 represent the historical extents of Morocco, and Tunisia). It is assumed that these countries transmission and do not reflect contemporary distributions have a negligible malaria burden; thus, they are excluded of risk (M. Craig, personal communication). Class 4 areas from the Sub-Saharan Africa malaria burden estimations. are the only areas where malaria still poses a risk, and its Second, areas with marginal risks of malaria transmission extent and transmission potential are determined by 198 | Robert W. Snow and Judy A. Omumbo Table 14.1 Populations at Risk During 2000 (thousands) Region Age 0­4 years Age 5­14 years Age 15 years Total North Africa (exclusion) 17,020 33,663 95,038 145,721 Southern Africa (classes 1­3: no malaria risk) 4,190 10,682 29,195 44,067 Rest of Africa (class 1: no malaria risk) 7,366 11,680 23,033 42,079 Southern Africa (class 4: stable malaria risk) 2,049 3,709 8,687 14,445 Rest of Africa (classes 2, 3: epidemic to low stable risk; 22,018 34,668 69,126 125,812 FCS 0.75; parasite prevalence 25%) Rest of Africa (class 4: stable endemic risk; 73,351 115,261 228,105 416,717 FCS 0.75; parasite prevalence 25%) Source: Compiled by authors. Note: The medium variant assumptions of changing fertility, mortality, and migration on population growth were selected for the year 2000. The database also provided the estimated proportion of the population age less than 5 years, 5­14 years, and 15 years or older and the crude birth rate (CBR). The CBR was used to estimate the approximate number of live births in 2000. The residents of Comoros, Réunion, and the Seychelles are assumed to be at negligible malaria risk and during the burden estimations formed part of the excluded Sub-Saharan Africa populations. Conversely it seems reasonable to assume that the islands of Cape Verde and São Tomé and Principe are subject to stable transmission; however, they have proved difficult to characterize at effective resolutions using the GIS malaria risk models, and their small contribution to populations at risk (less than 0.5 million people) have not been included. aggressive vector control. The combined malaria-risk and transmission (Barnes and Folb 2003; LSDI 2001). It will population models were used to estimate populations resi- never be possible to account for all local vector control dent in each malaria-risk class by using country-level data activities on a continent-wide map, but it is important to available from the World Population Prospects Population recognize that because current estimations of disease bur- Database (United Nations Population Division 2001). The den will be overestimated. summed extractions for the continent are shown in Other important factors that determine localized trans- table 14.1. mission intensity are currently not captured in the climate- The current distribution maps of P. falciparum depend driven models shown in figure 14.3. These include wide- entirely on the biotic effects of climate on transmission. spread use of insecticide-treated bednets, leading to a They fail to capture the more localized yet marked effects, decrease in local parasite transmission in only a few areas of such as urbanization and localized control. The Anopheles Sub-Saharan Africa; drug resistance, leading to increased gambiae complex is less prolific in urban areas because parasite transmission carriage; population displacement there are fewer appropriate breeding sites. People living in due to conflict or resettlement, altering transmission char- urban areas are on average 10 times less likely to receive an acteristics of newly settled areas; man-made ecological infective bite than their rural counterparts (Hay et al. 2000; changes, such as deforestation; and agricultural practices, Robert et al. 2003; Trape et al. 1992). In 2000, 38 percent of such as rice-field cultivation, that change the breeding site Africa's population were urban dwellers (United Nations availability for local mosquito vector populations. More Population Division 2001). The current estimations of detailed models of malaria risk in Mali and Kenya have populations at risk under different endemicity conditions included some of these factors on more recent maps (table 14.1) will overrepresent those in the higher-risk (Kleinschmidt et al. 2001; Snow et al. 1998). However, at classes. New satellite imagery makes it possible to map the present these more complex models are not available at a extent of land use, land cover, and urbanization, a refine- continental scale. ment important for subsequent iterations of the disease Finally, because population estimates in the United burden models for Sub-Saharan Africa (Hay et al. 2005). Nations Population Division's figures are based on estima- Several aggressive, and highly successful, vector control tions incorporating assumptions about fertility and mortal- programs in Africa have had significant effects upon trans- ity in a country, these totals may differ from the total popu- mission (Bradley 1991; Le Sueur, Sharp, and Appleton lations in each administrative unit reported by the country. 1993). More recently, such countries as South Africa and More refined microcensus data or models of population Mozambique have successfully reduced or eliminated projections below the national level are required for future transmission in areas that are currently classified as prone to mapping of disease burdens. Malaria | 199 DIRECT MALARIA MORTALITY with the terminal illness. The VA details are either reviewed by a panel of clinicians or subjected to diagnostic algo- Malaria is often difficult to diagnose on a purely clinical rithms. The sensitivity and specificity of VA diagnosis for basis. Fever is common to almost every infectious disease, malaria as a cause of death have been estimated in seven and the severe pathology caused by P. falciparum, such as aci- hospital-based validation studies in Africa. There is consid- dosis, anemia, and altered consciousness, are also erable variation in both the specificity (77 to 100 percent) complications of other infections. When a person is ill, and the sensitivity (45 to 75 percent) (Korenromp et al. demonstrating the presence of malaria infection increases 2003). The sensitivity of verbal autopsies for malaria will the likelihood that symptoms are directly due to the infec- depend on the intensity of malaria transmission: in high- tion, but the high prevalence of asymptomatic infections transmission areas, severe malaria is more likely to present as makes it difficult to exclude other diagnoses. Our earlier SMA than as CM (Snow et al. 1997a). SMA in a young child understanding of the pathophysiology of malaria derived is often difficult to distinguish from acute respiratory infec- from clinical descriptions among adults in Southeast Asia tions (ARI), although CM among older children is less and only recently have the mechanisms of death been more ambiguous. Similarly, the sensitivity and specificity of verbal precisely defined for pediatric African populations (Marsh autopsies for malaria will vary with the local spectrum of et al. 1995; Warrell 2002). Several detailed clinical studies in other diseases. ARI, acute gastroenteritis, and meningitis all African hospital settings have described the principal, some- share common clinical features with malaria, including times overlapping, routes to a fatal outcome. These include cough, difficulties in breathing, diarrhea, or cerebral dys- cerebral involvement from sequestered infection in the vas- function. Despite these limitations, the VA-diagnosed risks culature of the brain, metabolic disturbances, respiratory of malaria mortality represent our only source of contem- distress, and severe anemia. For epidemiological purposes it porary information on the direct, fatal consequences of is convenient to define two major syndromes, cerebral infection in areas with stable transmission in Africa. malaria (CM) and severe malarial anemia (SMA). CM is a The data presented in this chapter derive largely from the condition in which patients present in coma with several Burden of Malaria in Africa (BOMA) project, supported by underlying causes, ranging from a primarily neurological the Wellcome Trust, United Kingdom, and the Bill & condition to a systemic metabolic disturbance (Marsh and Melinda Gates Foundation. The BOMA project began in Snow 1999; Newton and Krishna 1998). Severe anemia is a 1998 and has completed a comprehensive search for pub- pathology of life-threatening malaria with a complex etiolo- lished and unpublished empirical measures of malaria mor- gy combining rapid hemolysis during acute infection or a bidity, disability, and death. The data include historical slow insidious process compounded by antimalarial drug accounts of mortality from colonial administration records resistance. SMA is a life-threatening condition in young chil- and detailed descriptions of recent DSS surveys. Details of dren and often warrants blood transfusion. the search methods and applications of the data are presented elsewhere (Snow, Korenromp, and Gouws 2004; Snow and Marsh 2002; Snow, Trape, and Marsh 2001). The Measuring Malaria-Specific Mortality estimates of malaria-specific mortality in childhood outside Africa presents an epidemiological challenge, because of southern Africa use only mortality reports based on a VA although it has high mortality rates, existing vital registra- within a DSS framework after 1989. This time restriction tion systems are not reliable or comprehensive. Most deaths was made because of recent temporal changes in pediatric occur outside the formal health service, and national gov- risks of mortality from malaria in Sub-Saharan Africa (see ernment systems of civil registration in Sub-Saharan Africa "Trends in Malaria Mortality" section). Where DSS and VA are incomplete. Epidemiologists interested in defining methods were used during community or household ran- malaria-specific mortality have established demographic domized intervention trials, only control communities were surveillance systems (DSS) of large populations (between included. In order to structure risks according to transmis- 20,000 and 100,000 people) to prospectively monitor popu- sion classes shown in table 14.1, DSS surveys of malaria- lation migration, births, and deaths (INDEPTH Network specific mortality were also limited only to those in which 2002; Smith and Morrow 1996). there was a congruent estimation of malaria transmission The attribution of causes of death during DSS surveys intensity. is often performed through a verbal autopsy (VA) interview Adult malaria-specific mortality data from DSS sites are with bereaved relatives about symptoms and signs associated rare, and VA methods have not been well developed nor 200 | Robert W. Snow and Judy A. Omumbo widely validated for the description of malaria deaths in this Saharan Africa. Data from southern Africa, however, also group. At Kilifi, Kenya, in the absence of obvious signs of underreport events and must be viewed as minimum esti- respiratory disease an algorithm was used based on detailed mates. Data have been extracted from reports provided dur- exclusion criteria (for example, well-defined deaths or ing subregional malaria control program meetings and hospital-diagnosed conditions) involving an acute febrile malaria deaths (reported only for all age groups combined) event (R. W. Snow, unpublished data). In Tanzania a similar and expressed per projected population estimates for approach was used (Y. Hemed, personal communication). administrative areas of known malaria risk (figure 14.3). Both approaches are likely to overestimate the number of Data were available for the three areas of KwaZulu-Natal adult deaths from malaria in endemic areas. Owing to the Province in South Africa, seven areas of Botswana, four paucity of empirical data in this age group, these two con- areas of Namibia, and eight areas of Zimbabwe. temporary estimates have been included with other sources The spatial coordinates of the study location were used of historical data on adult malaria mortality from detailed, to identify a time and location-specific estimate of infection preindependence civil registration systems when causes of prevalence for the geographical areas covered by the death were investigated by medical personnel from circum- mortality data. Data were identified primarily through the scribed, censused populations. It seems reasonable to MARA/ARMA database (http://www.MARA/ARMA.or.za) assume that under stable transmission conditions adult or correspondence with authors of original data on the mortality from malaria is unlikely to have changed much disease burden. The estimates of malaria risk have been used over time, as deaths would normally occur in individuals to provide empirical ecological categorization for each who, for some ill-defined reasons, failed to acquire func- data point in accordance with the cartography of risk tional immunity in childhood or lost their functional described in figure 14.3 and table 14.1. Median and immune response. The intention has been to attempt, where interquartile ranges (IQR) have been used to describe the possible, to fill an important data gap using imperfect data, risk data per malaria risk class (table 14.2), and the and therefore the risks reported must be interpreted with estimated numbers of deaths directly attributed to caution. P. falciparum in 2000 are shown in table 14.3 according to Civil and vital registration systems in southern Africa are populations at various continental risk classes shown in more comprehensive than those in other parts of Sub- table 14.1. Table 14.2 Median and Interquartile Ranges of Malaria-Specific Mortality Estimates per 1,000 People per Year Region Age 0­4 years Age 5­14 years Age 15 years Bibliographic sources Southern Africa (malaria 0.13 [0.08­0.21] Diseko 1994; Kamwi 1998; Piotti 1997; South African risk areas; class 4) (study sites 22) Medical Research Council 1998, unpublished data Rest of Africa (low, stable 2.62 (287/109,412 0.94 (158/167,598 0.71 (265/372,777 Charlwood et al. 2001; Government of epidemic risk; classes 2, 3; PYO) (study sites 3) PYO) (study sites 2) PYO) (study sites 2) Tanzania 1997 FCS 0.75; parasite prevalence 25%) Rest of Africa (stable 9.33 [7.38­14.57] 1.58 [0.66­2.77] 0.6 [0.37­0.94] Delacollette and Barutwanayo 1993; Ghana VAST 1993; endemic risk; class 4; (study sites 12) (study sites 10) (study sites 15) Jaffar et al. 1997; Snow, Mung'ala et al. 1994; FCS 0.75; parasite Trape et al. 1998; Barnish et al. 1993; Premji et al. 1997; prevalence 25%) Government of Tanzania 1997; Salum et al. 1994; D'Allesandro et al. 1995; Pasha et al. 2003; Government of Colony of Gold Coast 1912­48; Colbourne and Edington 1956; Government of the Colony and Protectorate of Kenya 1935; Government of Colony and Protectorate of Nigeria 1934­35; Bruce-Chwatt 1952; Government of Colony and Protectorate of Sierra Leone 1913­17 Source: Compiled by authors. For low, stable endemic risks for "rest of Africa," G. D. Shanks and R. W. Snow provided unpublished data in Kenya; additional information from the Ghana VAST project were provided by the late Nicola Dollimare and Gilly Maude; unpublished VA data reconstructed by R. W. Snow to provide estimates of malaria-specific mortality in the study described by Snow, Mung'ala et al. 1994; additional information provided by Brian Greenwood for the Barnish et al. 1993 study. Note: First numbers in each entry represent the median; numbers in square brackets represent the IQR; PYO person-years of observation. For southern Africa, the figures in the first column include all age groups. Malaria | 201 Table 14.3 Estimated Numbers of Malaria-Specific Deaths and Interquartile Ranges during 2000 Region Age 0­4 years Age 5­14 years Age 15 years Total Southern Africa (malaria risk areas; class 4) 1,878 Rest of Africa (low, stable epidemic risk; 57,688 32,588 49,079 139,355 classes 2 & 3; FCS 0.75; parasite prevalence 25%) Rest of Africa (stable endemic risk; class 4; 684,364 182,113 136,863 1,003,340 FCS 0.75; parasite prevalence 25%) [541,330­1,068,723] [76,072­319,274] [84,399­214,419] [701,801­1,602,416] Source: Compiled by authors. Note: The total for southern Africa includes all age groups. Numbers in square brackets represent the IQR. Empirical Estimates of Malaria-Specific Mortality by Age mortality rate was 0.94 per 1,000 per year (two studies); and Twelve independent estimates of malaria mortality among for adults age 15 years and over the rate was 0.71 per 1,000 children under five years living under conditions of stable per year (two studies). transmission were used from DSS sites in Burundi, The An ongoing rural DSS site at Agincourt, South Africa Gambia, Ghana, Kenya, Senegal, Sierra Leone, and Tanzania. (Kahn et al. 1999), located at the fringes of malaria risk The median malaria-specific mortality rate among these (Brink 1958), also provided additional data. A malaria- communities was 9.33 per 1,000 children per year and rep- specific mortality rate of 0.065 per 1,000 per year (2 out of resented 28.2 percent of all mortality among children under 216 deaths) was measured among children under five years five. Under similar transmission conditions the median between 1992 and 1995. Of the 785 deaths recorded in the malaria-specific mortality among older children, age 5 to population older than five years of age, only one was attrib- 14 years, was 1.58 per 1,000 per year (from 10 study sites), uted to malaria. This single estimate is an inadequate rep- representing 52.2 percent of all deaths in this age group. resentation of the malaria mortality across southern Africa. Among adult populations surveyed as part of colonial The civil registration data in malaria-risk districts of south- administration civil registration systems or recent DSS sites ern Africa suggest that the median estimate of malaria- in stable endemic areas, the median malaria mortality rates specific mortality among the entire population in these were 0.6 per 1,000 per year (15 reports), or 6 percent of all areas was 0.13 per 1,000 per year. Although these estimates deaths among populations age 15 years or older. represent minimal approximations of the true mortality There are few DSS or colonial administration civil regis- burden, they fall within the ranges of mortality described tration data from low, stable endemic or epidemic condi- for the single DSS site (Kahn et al. 1999) and those sites tions (parasite rate less than 25 percent; classes 2 and 3) out- described in the more detailed analysis of malaria mortali- side southern Africa. Approximations to the DSS from the ty data from civil registers in two districts of KwaZulu- populations of a tea estate in the highlands of Kenya Natal Province between 1996 and 1999: 0.02 to 0.52 per (Shanks and Snow, unpublished data) and a settled refugee 1,000 people per year (Tsoka, Sharp, and Kleinschmidt camp in Sudan (Charlwood et al. 2001) provided data on 2002). malaria-specific mortality rates among censused childhood Overall malaria-specific mortality in children is approxi- populations at low risk of malaria infection. In addition, the mately 3.5 times higher in areas of stable endemic transmis- one true DSS site at Hai district in Tanzania provided a VA sion than in areas of low intensity, stable, or epidemic-prone estimate of malaria mortality (Government of Tanzania malaria in Sub-Saharan Africa, excluding southern Africa. 1997). All three studies covered periods after 1990. Only the Mortality declines rapidly with increasing age, and this is Kenyan and Tanzanian studies provided information on especially striking under conditions of stable endemic trans- older children and adults. Rather than being used to provide mission. The mortality rates for all ages from P. falciparum in a median estimate from the limited data, the combined southern Africa are considerably lower than those described person-years of observation (PYO) and malaria deaths have for the rest of Africa, reflecting a low risk of infection com- been used to define direct risks of death from malaria bined with effective control. Even with the use of all the tools under these transmission conditions. For children under described earlier in this chapter, approximately 1.14 million five years the overall mortality rate was 2.62 per 1,000 per people might have died in Sub-Saharan Africa as a direct year (three studies); for children age 5 to 14 years the consequence of infection with P. falciparum in 2000. 202 | Robert W. Snow and Judy A. Omumbo CONSEQUENTIAL MORTALITY and malaria status of 167 admissions of children to an emer- gency ward at the Mama Yemo Hospital in Kinshasa, the The consequences of disease that are related to the clinical Democratic Republic of the Congo. The authors propose an event include the consequences of clinical management, unadjusted odds ratio for acquired HIV infection of 3.5 for such as the immediate effects of adverse drug reactions or malaria patients between 1 month and 12 years of age trans- the longer-term effects of HIV-acquired infection through fused once, 21.5 for those transfused twice, and 43.0 for those blood transfusion. Nonintervention-related consequences transfused three times during a single admission. In a study of clinical events also include the short- and long-term of transfusion practices in 1994 at six government hospitals residual impairments resulting from cerebral malaria. in Kenya, Moore and colleagues (2001) calculated a 2 percent risk of transmission of HIV antibody positive blood from screened donations through blood transfusion to HIV- Adverse Drug Reactions to Commonly Used negative patients. This study did not include the risk from Antimalarial Drugs pre-seroconversion donations, nor did it allow for the sensi- Drugs used to manage malarial fevers have adverse effects. tivity of test kits or the use of unscreened blood. The proba- We can assume that most of the severe adverse drug-related bility that an HIV antibody negative unit of blood is HIV events are described as direct mortality during DSS VA infected has been estimated to be between 0.5 and 1.1 percent studies, as they are likely to occur close to the febrile event. (Savarit et al. 1992). The probability of seroconversion fol- They are mentioned here, however, to emphasize that fevers lowing HIV-contaminated blood is assumed to be 96 percent are common in Africa and the use of antimalarial drugs is (Colebunders et al. 1991). From estimates of hospitalized prolific. Adverse drug reactions (ADR) are likely to increase patients with severe malaria anemia, transfusion rates, and with the greater use of new, more complex drug combina- relative risks of seroconversion from HIV-infected blood tions and agents, although the increase of ADR will proba- donation, it has been estimated that between 5,000 and 8,000 bly never surpass the alternative of using ineffective drugs. children might become infected with HIV each year as a There have been too few epidemiological studies of the consequence of poor management of their malaria in the human toxicity of many antimalarial drugs among African hospital (Snow et al. 2003). populations repeatedly exposed to these compounds. Most data derive from an examination of chemoprophylactic drug use among nonimmune travelers (Philips-Howard Neurological Disability Associated with Severe Malaria and Bjorkman 1990; Philips-Howard and West 1990). The The case-fatality rate of cerebral malaria in most hospital majority of adverse reactions due to sulfonamides and settings is high, often over 30 percent (Newton and Krishna 4-aminoquinolines are idiosyncratic. Severe adverse reac- 1998). Prolonged coma and seizures are associated with tions to the commonly available antimalarial drugs in neurological impairment in survivors. The immediate and Africa, when used as recommended, include severe cuta- prolonged sequelae associated with CM among African chil- neous reactions (Stevens-Johnson and Lyell's disease syn- dren includes hemiparesis, quadriparesis or severe deficit, dromes), aplastic anemia, severe neutropenia, thrombocy- hearing and visual impairments, speech and language and topenia, keratopathy, agranulocytosis, and hepatic failure nonverbal construction difficulties, behavioral problems, (Reynolds 1993). It has been estimated that 2,350 deaths and epilepsy (Mung'ala-Odera, Snow, and Newton 2004). were probably caused by treatment rather than the disease These impairments are estimated to occur in about 4,000 itself in a single year among the children outside of south- children each year in Sub-Saharan Africa. ern Africa (Snow et al. 2003). Those with severe deficits have a higher mortality risk soon after the disease event. For less dramatic impairments, such as epilepsy, increased mortality rates have been Anemia, Transfusion, and HIV described (Coleman, Loppy, and Walraven 2002; Jillek-All Severe anemia is a common feature of children hospitalized and Rwiza 1992; Snow, Mung'ala, et al. 1994). In Western with complicated malaria. Transfusion is a common pedi- countries, the risk of premature mortality could be as high atric practice in Africa and adherence to guidelines for trans- as two to three times that described in age-comparable fusion are often poor (English et al. 2004; Lackritz et al. groups without epilepsy (Cockerell et al. 1994; Hauser, 1992). Greenberg and colleagues (1988) examined the HIV Annegers, and Elveback 1980; Zielinski 1974), but in Africa Malaria | 203 the risk could be as high as nine times (Coleman, Loppy, and Korenromp, and Gouws 2004). The model regards P. falci- Walraven 2002). This is likely to be caused by poorly man- parum infection as a risk factor for all-cause mortality rather aged epilepsy resulting in status epilepticus or accidents, than attempting to directly estimate malaria's contribution such as drowning or burns. to all-cause mortality (one cause, one death). Malaria during Pregnancy INDIRECT MORTALITY Despite a poor understanding of the precise mechanisms of Indirect consequences of P. falciparum infection include pathology (Menendez 1995), the morbid outcomes of anemia (unless anemia is linked to acute high-density malaria infection during pregnancy have been well parasitemia as a direct cause), low birthweight, growth described (Brabin 1983; Guyatt and Snow 2001a, 2001b; retardation, or undernutrition. In addition, malaria infec- Steketee, Wirima, Hightower et al. 1996; Steketee et al. tion can increase the severity of other comorbid infectious 2001). In endemic settings in Africa, pregnant women expe- diseases through immune suppression or enhanced invasive rience relatively little malaria-specific morbidity (for exam- capacities across physical barriers to infection (for example, ple, fever) but do have increased risk of infection and higher blood and tissue). Previous approaches to the global burden density parasitemia leading to anemia and placental seques- of disease have assumed that each death must be attributed tration of the parasite. These effects operate across a broader to a single cause and can be fitted into the fixed disease-mix range of endemicities used to describe morbid and fatal matrix of all causes (Murray and Lopez 1997). risks among nonpregnant populations (that is, areas cov- During randomized controlled intervention trials aimed ered by classes 3 and 4 shown in figure 14.1). Maternal at reducing the incidence of infection (but not 100 percent anemia has been shown to be an important contributor to protective), the all-cause mortality of children is often maternal mortality with a relative risk for mortality of 1.35 reduced more than would be attributed by VA diagnosis of for moderate anemia and 3.51 for severe anemia (Brabin, malaria. For example, in Kilifi the proportion of deaths of Hakimi, and Pelletier 2001). Brabin and colleagues further children under five years attributed to malaria by VA was 34 estimate that malaria contributes to maternal anemia and percent (R. W. Snow, unpublished data). During a random- that 9 percent of anemia-associated maternal mortality can ized controlled trial of insecticide-treated bednets in the be attributed to malaria in Sub-Saharan Africa; this figure same area, the incidence of malaria infection was reduced by predicts approximately 5,300 maternal deaths annually 50 percent (Snow et al. 1996), which was sufficient to reduce from malaria anemia in areas of Sub-Saharan Africa outside all-cause mortality by 33 percent (Nevill et al. 1996). More of southern Africa classified as classes 3 and 4. dramatically, in The Gambia, insecticide-treated bednets Prematurity and low birthweight (less than 2,500 grams) reduced all-cause mortality by over 60 percent, and yet the are associated with maternal malaria, including the contri- VA-diagnosed contribution of malaria to all-cause mortali- bution from both malaria-associated maternal anemia and ty among control populations was only 16 percent (Alonso placental infection. The contribution of malaria during et al. 1993). This has led some to speculate that malaria pregnancy to low birthweight and subsequent mortality in infection is a contributor to broad causes of mortality the first year of life has been estimated to range from 3 to beyond the direct fatal consequences of infection 8 percent of infant mortality (Greenwood et al. 1992; (Molineaux 1997). Steketee, Wirima, Hightower et al. 1996; Steketee et al. Attempts have been made to distinguish between factors 2001). If this range had been applied to the expected num- that affect the direct outcomes of malaria infection and the bers of live births in 2000 among populations in low effects of malaria infection on the outcome of other health endemic, epidemic prone, and endemic malaria areas of burdens. This is particularly important when the interaction Sub-Saharan Africa outside of southern Africa, there may between malaria and HIV or undernutrition is considered. have been between 71,000 and 190,000 infant deaths indi- Especially important is the role malaria infection plays as a rectly attributable to malaria in pregnancy. risk for the extended, indirect consequences of infection In very low endemic settings or areas where malaria is on the public health burden posed by P. falciparum. A epidemic prone and in southern Africa, malaria in nonim- recent model relates infection prevalence to all-cause pedi- mune pregnant women can be devastating and lead to atric mortality outcomes in Sub-Saharan Africa (Snow, maternal death and abortion. For example, in urban 204 | Robert W. Snow and Judy A. Omumbo Mozambique, 15.5 percent of all maternal deaths in one specific nutrient deficiency, and only a few have examined hospital over a five-year period were attributed directly to the role of malaria. One striking feature of the global distri- malaria (Granja et al. 1998). In an epidemic-prone setting of bution of anthropometric markers of undernutrition is its Ethiopia, maternal malaria carried an approximately eight- congruence with the distribution of endemic malaria. fold increased risk of abortion (R. Newman, unpublished Although P. falciparum malaria and malnutrition are both data). Similarly, in a low-endemicity setting in Southeast highly prevalent in Sub-Saharan Africa, the existence of a Asia even single malaria infections were associated with synergistic interaction has not been well established. increased risk of low birthweight (Luxemburger et al. 2001). Evidence from intervention trials aimed at reducing the fre- quency of new infections suggests that malaria infection might have some indirect effects upon the generalized nutri- Malaria and Anemia tional status of African children. A study in Nigeria on the Anemia among African children is a hematological state use of chemoprophylaxis in the treatment of malaria in chil- determined by combinations of nutritional deficiencies dren showed a reduction in the incidence of infection and (iron, folic acid, other micronutrients, and protein-calorie clinical attacks that was accompanied by a reduction in the malnutrition), iron loss through helminth infection, red cell incidence of malnutrition (Bradley-Moore et al. 1985). destruction, red cell production decreased by infectious dis- Improved growth among young children has more recently eases, and the genetic constitution of red cell hemoglobin been demonstrated in The Gambia and Kenya in studies (Menendez, Fleming, and Alonso 2000; Nussenblatt and comparing those protected by insecticide-treated bednets Semba 2002). Malaria has long been recognized as a major with those left unprotected (D'Allessando et al. 1995; Snow contributor to anemia in children, reducing hemoglobin et al. 1997b; Ter Kuile et al. 2003). Despite the biological concentrations through several mechanisms. The primary plausibility of synergism between infection and growth, the mechanism increases rates of destruction and removal of red precise relationship between undernutrition and severe blood cells and decreases the rate of erythrocyte production malaria continues to be difficult to quantify empirically in the bone marrow. Other mechanisms are associated with within disease burden frameworks. acute clinical states (for example, hemolysis or cytokine dis- turbances), whereas chronic or repeated infections are more Malaria and HIV Interactions likely to involve dyserythropoiesis (Menendez, Fleming and Alonso 2000). Data from randomized controlled trials of In Sub-Saharan Africa the HIV epidemic has been superim- malaria-specific interventions to reduce the incidence of new posed on the long-standing malaria pandemic. The wide infections through insecticide-treated bednets or the preva- geographical overlap and the concurrent high prevalence of lence of blood-stage infections through chemoprophylaxis both HIV and malaria mean that even modest interactions or intermittent presumptive treatment suggest a halving of could substantially affect public health among populations anemia risks through intervention (Korenromp et al. 2004). exposed to both (Chandromohan and Greenwood 1998). This substantial reduction in the risks of anemia demon- Studies during the 1990s observed that malaria infection strates the importance of P. falciparum in maintaining the was more common and of higher parasite density in HIV- poor hematological health of children. This finding, howev- positive than in HIV-negative pregnant women in a range of er, cannot be extrapolated to subsequent morbid or fatal con- malaria endemic settings, and in women of all gravidity, and sequences, because few detailed prospective studies focus on that those who have had multiple pregnancies were most describing the effects of reduced hemoglobin concentrations affected (Parise et al. 1998; Shulman 1999; Steketee, Wirima, on health outcomes. Bloland et al. 1996; van Eijk, Ayisi, and ter Kuile 2001, p. 405; van Eijk et al. 2001; Verhoeff et al. 1999). Additionally, two longitudinal cohort studies in Uganda and Kenya and one Malaria and Undernutrition hospital-based case-control study in Uganda have demon- It has been postulated that nutritional status is related to the strated that HIV-infection approximately doubles the risk of threats posed by infection and disease and to the role of malaria parasitemia and clinical malaria in nonpregnant infectious disease in perpetuating undernutrition (Pelletier, adults, and that increasing HIV-immunosuppression is Frongillo, and Habicht 1993; Pelletier et al. 1995). Most associated with higher-density parasitemias (Francesconi studies have generally focused on severe malnutrition and et al. 2001; French et al. 2001; Whitworth et al. 2000). Thus, Malaria | 205 some evidence shows that HIV infection increases the inci- or equal to 25 percent. In the regression model, mortality dence and severity of clinical malaria in adults. increased significantly with parasite prevalence, but this In order to define the indirect consequences of P. falci- effect leveled off at higher prevalence rates. The model sug- parum it is important to know whether the high intensity of gested that, in rural DSS sites throughout Sub-Saharan exposure to malaria infections increases the rate of pro- Africa, all-cause mortality increases by more than twofold gression of HIV disease in Africa. In a recent study in (25­30 deaths per 1,000 children under five years old) over Malawi, HIV blood viral levels were found to be seven the prevalences of malaria infection covered by the DSS sites, times higher in HIV-infected adults with acute uncompli- and parasite prevalence explained 64 percent of the variation cated malaria than in HIV-infected blood donors without between sites in all-cause under-five mortality. By contrast, malaria (Hoffman et al. 1999). As with other acute infec- the direct estimation of malaria-specific mortality presented tions, the increased viral burden was reversed by effective earlier for children living under stable endemic conditions malaria therapy (Hoffman et al. 1999). These findings are was only 28.2 percent. consistent with in vitro laboratory studies in which HIV-1 These comparisons of direct versus indirect contribu- replication was increased 10-fold to 100-fold in peripheral tions of malaria infection to child survival must be viewed blood mononuclear cells exposed to soluble malaria anti- with some caution. Ecological analyses are constrained by gens or malaria pigment (Xiao et al. 1998). multiple confounders, notably, influences of socioeconomic The evidence on the potentiation of malaria by HIV or status, access to and effectiveness of health services, preva- HIV by malaria is currently insufficient to quantify the lence of HIV, exposure to other parasitic diseases, nutrition, specific HIV-malaria interaction risks for inclusion in and the genetic makeup of the population. Still, such an malaria burden estimates. approach to estimating malaria's contribution to child mor- tality in Sub-Saharan Africa suggests that the influence of P. falciparum infection is greater than that described by the The Combined Indirect Effects of Malaria direct, single-cause attribution of childhood deaths. The on All-Cause Mortality effects of malaria during pregnancy on low birthweight and The effects of malaria infection on low birthweight seem to a lesser extent the nutritional and hematological influ- conclusive; the overall role of malaria infection on anemia is ences of malaria infection may explain the difference clear, but its extrapolation to indirect mortality is difficult; between direct and indirect estimates. the effects of infection on undernutrition and HIV is at present less conclusive and impossible to enumerate. It seems reasonable to assume that these additional conse- MALARIA AND POVERTY quential or indirect risks contribute to all-cause mortality beyond that described from estimates of the causes of death Recent global, cross-country regression analysis of malaria directly attributed to malaria. risk and gross domestic product have found the disease to be To examine these conclusions and assumptions further, a significant influence on long-term economic growth. As data on all-cause mortality of children under five from DSS much as half the gross domestic product is lost in highly studies undertaken across a broad range of malaria transmis- endemic areas over 25 years as a direct result of malaria sion settings in Sub-Saharan Africa were analyzed against the (Gallup and Sachs 2001; McCarthy and Wu 2000). Although prevalence of P. falciparum infection at each site. Weighted persuasive to international donors, such a macroeconomic least-squares regression was used to model the contiguous analysis fails to identify the mechanisms of these economic relationships between all-cause mortality and parasite preva- losses (Malaney, Speilman, and Sachs 2004). It seems reason- lence rates, allowing for the square of parasite prevalence (for able to assume that national levels of economic loss are a possible saturation of parasite prevalence), timing, location, composite of economic burdens at the household level. and the sampling precision of each study (Snow, Korenromp, Shepard and colleagues (1991) estimated the household cost and Gouws 2004). The unadjusted median all-cause child of malaria in Burkina Faso, Chad, the Republic of Congo, and mortality rate for low prevalence areas of childhood infection Rwanda. The authors concluded that a case of malaria in (less than 25 percent) was 10.9 per year per 1,000 children Africa cost US$9.84 in 1987, of which US$1.83 was direct and under five (IQR 7.8­17.6). This rose dramatically to 39.1 per US$8.01 was accrued indirectly as a result of forgone income year per 1,000 children (IQR 32.8­52.2) among populations associated with malaria morbidity and mortality. The total exposed to childhood parasite prevalence risks greater than estimated cost of US$0.8 billion represents 0.6 percent of the 206 | Robert W. Snow and Judy A. Omumbo gross domestic product of the economies in Sub-Saharan Figure 14.4 Malaria-Specific and All-Cause Mortality Africa. Thus, microeconomic analyses of household eco- Estimates per Year for Children under Five nomic burden posed by malaria are only a fraction of macro- (per 1,000) economic analyses of national economic loss. Malaney, a. Malaria-Specific Mortality Speilman, and Sachs (2004) suggest that this difference might 20 arise either because of inadequate methodologies or because 18 the macroeconomic national burden encompasses house- hold financial costs and ill-defined financial "externalities," 15 such as trade, tourism, and investment. 13 Whether malaria drives household or national poverty deaths 10 requires better definition. However, the converse is malaria 8 irrefutable: poor people are less able to prevent infection or no. afford effective disease management. In Africa, malaria is 5 largely a disease of the rural populations, and often these 3 communities are home to some of the poorest of the poor 0 in Africa. There is increasing evidence that strategies pro- 8 20 11 1960 1960­89 1990 moted to prevent infection, such as insecticide-treated bed- N and years nets, are not reaching the poor when cost-retrieval is part of the strategy. The recent Kenyan Demographic and Health b. All-Cause Mortality 80 Survey showed that less than 7 percent of children described as living in households at the lowest wealth index quartile 70 sleep under an insecticide-treated bednet compared with 35 60 percent of children in the top wealth quartile households 50 (http://www.measuredhs.org/). Similar findings have been deaths reported for Uganda (Mugisha and Arinatwe 2003). In 40 Tanzania, poor children were less likely to receive antimalar- all-cause 30 no. ials when febrile than children from wealthier families 20 (Schellenberg et al. 2003). A household survey in Malawi 10 focused on low-income households whose mean annual income was US$115 and where the costs of malaria preven- 0 8 20 11 tion and treatment represented about 20 percent of annual 1960 1960­89 1990 income (Ettling et al. 1994). N and years Source: Snow, Trape, and Marsh 2001. Note: Central horizontal lines median quartile range; box width and T median and upper/lower limits of mortality estimates, respectively. Single outlier is more than three times TRENDS IN MALARIA MORTALITY the box width. One of the earlier applications of the BOMA project was to examine long-term changes in the estimates of malaria mor- a median estimate of 10.2 per 1,000 children under five per tality in Sub-Saharan Africa (Snow, Trape, and Marsh 2001). year, similar to that described in the preindependence sur- A comparison of preindependence malaria-specific and all- veillance data (figure 14.4a). In contrast, the annualized cause childhood mortality data derived from circumscribed risks of all-cause mortality recorded at the same sites colonial medical records and civil registration with early demonstrate a continued decline over the three time periods DSS data before 1990 and contemporary DSS data post- (figure 14.4b), and overall childhood mortality during the 1990 is shown in figure 14.4. The period 1960­89 was char- early 1990s was 34 percent lower than the median estimates acterized by a median malaria-specific mortality of 7.8 per of all-cause mortality recorded during the preindependence 1,000 children under five per year, a decline of 18 percent surveys. The net result has been a rise in the proportion of from the estimates recorded before 1960 (9.5 per 1,000 chil- all deaths attributed to malaria: 18 percent before 1960, dren under five per year). However, the 11 studies of child- 12 percent between 1960 and 1989, rising to 30 percent hood malaria-specific mortality during the 1990s provided during the 1990s. Malaria | 207 The BOMA data were reanalyzed to define more precisely temperature do not support a trend in favor of increased the recent trends in malaria-specific mortality estimates malaria transmission (Hay et al. 2002). For many years from DSS sites among children under five years, allowing for chloroquine provided a cheap, effective, and easily available variation in the VA performance, malaria endemicity, and treatment. The past 20 years have seen a precipitous decline region (Korenromp et al. 2003). In West African DSS sites, in the efficacy of chloroquine across Africa (EANMAT 2003; malaria mortality was on average 7.8 per 1,000 child-years Talisuna, Bloland, and D'Alessandro 2004), and this repre- throughout the 1980s and 1990s without a significant sents the most likely factor contributing to the change in change over time. Although conclusive evidence of tempo- malaria-specific childhood mortality. Such an explanation ral increases in malaria-specific mortality was observed would also be entirely consistent with observations of within a series of DSS sites in Senegal (Trape et al. 1998). In increased hospitalization and morbidity in other parts of East and southern African DSS sites the estimated malaria Africa (Asindi et al. 1993; Greenberg et al. 1989; Shanks et al. mortality increased from 6.5 per 1,000 child-years between 2000) and the sharp rises in malaria morbidity associated 1982 and 1990 to 11.9 per 1,000 child-years between 1990 with declining first-line therapy in KwaZulu-Natal Province and 1998. Nonmalaria mortality, in contrast, decreased over of South Africa and its reversal following the introduction of time in both regions. Figure 14.5 summarizes the combined effective artemisinin-based combination therapies (Barnes Sub-Saharan Africa data for 1982­98, allowing for parame- and Folb 2003). ters of region, parasite prevalence, and VA performance. There are several possible explanations for the observed trends in malaria mortality in Sub-Saharan Africa over the CONCLUSION last 20 years, such as declining household wealth linked to a changing health sector based on cost retrieval and the gen- During 2000 approximately 1.14 million people may have eral deterioration in the quality of clinical care. Although died as a direct result of infection with P. falciparum. Eighty- access to and quality of health care is undoubtedly subopti- eight percent of these deaths would have occurred in areas mal in many areas, this has not had a similar impact on non- of stable endemic malaria and the majority of these would malaria childhood mortality. Global warming has been have been among young children. The empirical evidence debated as a general cause of expanding malaria risk and for indirect or consequential mortality may explain an addi- thereby increasing malaria-specific mortality in several parts tional 10 percent of mortality directly attributed to malaria of Africa. However, this seems an unlikely explanation for infection. This would be consistent with observations made the areas included in figures 14.4a and 14.5, as long-term during randomized controlled trials that suggest that reduc- data for areas likely to be affected by changes in ambient ing the risks of infection in a community has an impact beyond what might readily be described by verbal autopsy as Figure 14.5 Malaria-Specific, Nonmalaria, and All-Cause direct malaria mortality. Recent ecological analyses of all- Mortality Rates among Children under Five cause pediatric mortality suggest that the difference between (per 1,000 per year from 28 DSS sites) direct and indirect attribution of malaria as cause of death 60 might be substantially higher than 10 percent. p 0.002 p 0.001 The Roll Back Malaria movement proposes to halve 50 malaria mortality by the year 2010 (http://www.rbm.int). 40 p 0.0022 This goal has been set even though existing, affordable ther- rate apeutics are rapidly failing, health service provision is break- 30 ing down, there are no immediate prospects of widespread mortality 20 vaccination, and poverty continues to afflict most endemic p 0.011 10 countries. There are strong reasons to believe that over the past 15 years malaria-specific mortality has risen and now 0 malaria nonmalaria all cause malaria (percent) accounts for an increasing proportion of overall childhood mortality. The starting point for new efforts to "roll back 1982­89 1990­98 malaria" is not a level playing field but a mortality burden Source: Adapted from Korenromp et al. 2003. that has returned to levels described before Africa gained Note: Data analyzed by least-squares linear regression to allow for interactions of determinants of VA-adjusted mortality, the square of parasite prevalence, and region. independence. 208 | Robert W. Snow and Judy A. Omumbo One million deaths due to malaria each year in Africa res- NOTE onate with earlier claims of a similar figure proposed as far back as the 1950s (Bruce-Chwatt 1952; Greenwood 1990; This review was made possible because of a large contribu- Schwartlander 1997; Sturchler 1989). These estimations are tion over the years from malaria epidemiology colleagues, hard to comprehend without a methodological framework notably Marlies Craig, Uwe Deichmann, Simon Hay, Eline or empirical evidence to support them. How much further Korenromp, Claire Mackintosh, Kevin Marsh, Charles forward are we in estimating malaria's contribution to the Newton, Dennis Shanks, Rick Steketee, and Jean-François mortality burden in Sub-Saharan Africa? The estimates pro- Trape. The Burden of Malaria in Africa Project is principally vided in this chapter continue to be driven by informed supported by the Wellcome Trust, United Kingdom (project approximations, in part because of the paucity of reliable number 058992), with additional support from the Bill & and accurate data, but also due to the inherent difficulties of Melinda Gates Foundation (project number 17408); the unique diagnosis. The estimates, however, have been made Disease Control Priorities Project, World Bank; and the from empirical epidemiological measures of mortality risks, Kenyan Medical Research Institute. Robert W. Snow is sup- structured according to age and malaria transmission. The ported by the Wellcome Trust, United Kingdom, as a senior method is similar to one previously used to define the research fellow. malaria burden in Sub-Saharan Africa (Snow et al. 1999) but takes into account new data, refined endemicity classifications, broader health consequences, and the tempo- REFERENCES ral effects associated with changing antimalarial drug sensi- tivity. This data-driven method allows for a more transpar- Alonso, P. L., S. W. Lindsay, J. R. M. Armstrong-Schellenberg, K. Keita, P. Gomez, F. C. Shenton, A. G. Hill, et al. 1993. "A Malaria Control ent review of the evidence. Trial Using Insecticide-Treated Bed Nets and Targeted Chemo- Such reviews of available evidence reveal as much about prophylaxis in a Rural Area of The Gambia, West Africa." Transactions of the Royal Society of Tropical Medicine and Hygiene 87 (Suppl. 2): what we do know as about what we do not know. Many 37­44. aspects of the malaria burden in Sub-Saharan Africa Asindi, A. A., E. E. Ekanem, E. O. Ibia, and M. A. Nwangwa.1993. "Upsurge require further empirical investigation or modeling. In the of Malaria-Related Convulsions in a Paediatric Emergency Room in continued absence of wide-area, reliable coverage of cause- Nigeria. Consequence of Emergence of Chloroquine-Resistant Plasmodium falciparum." Tropical and Geographic Medicine 45: 110­13. specific mortality data there is a need for increased research Barnes, K., and P. Folb. 2003. "The Role of Artemisinin-Based to help in an understanding of the spatial determinants of Combination Therapy in Malaria Management." In Reducing Malaria's risk. The assumption that urban versus rural, East versus Burden: Evidence of Effectiveness for Decision Makers, ed. C. Murphy, K. Ringheim, S. Woldehanna, and J. Volmink, 25­31. Washington, DC: West African, or poor versus affluent populations experi- Global Health Council. ence similar risks of poor health outcomes from malaria Barnish, G., G. H. Maude, M. J. Bockarie, T. A. Eggelte, and B. M. infection is clearly an oversimplification of complex inter- Greenwood. 1993. "The Epidemiology of Malaria in Southern Sierra Leone." Parasitologia 35: 1­4. actions. The precise cartography of risk requires improved Beier, J. C., G. F. Killeen, and J. I. Githure. 1999. "Entomologic Inoculation population distribution maps. This is now being addressed Rates and Plasmodium falciparum Malaria Prevalence in Africa." at a global scale through a new initiative, the Malaria Atlas American Journal of Tropical Medicine and Hygiene 61: 109­13. Project, funded by the Wellcome Trust, United Kingdom Brabin, B. J. 1983. "An Analysis of Malaria in Pregnancy in Africa." Bulletin of the World Health Organization 61: 1005­16. (http://www.map.ox.ac.uk). The emphasis within the liter- Brabin, B. J., M. Hakimi, and D. Pelletier. 2001."An Analysis of Anemia and ature on measuring childhood risks of disease and death Pregnancy-Related Maternal Mortality." Journal of Nutrition 131 from malaria would seem justified from the current under- (Suppl.): 604S­615S. standing of immunity, but there is comparative ignorance Bradley, D. J. 1991. "Morbidity and Mortality at Pare-Taveta, Kenya and Tanzania, 1954­66: The Effects of a Period of Malaria Control." In about the true public health consequences of malaria Disease and Mortality in Sub-Saharan Africa, ed. R. G. Feachem and among older children and adults. The impact of infection D. T. Jamison, 248­63. New York: Oxford University Press. and disease on undernutrition, anemia, and HIV remain at Bradley-Moore, A. M., B. M. Greenwood, A. K. Bradley, B. R. Kirkwood, and H. M. Gilles. 1985. "Malaria Prophylaxis with Chloroquine in best speculative and require further investigation before Young Nigerian Children. III. Its Effect on Nutrition." Annals of these composite risks can be defined. The most significant Tropical Medicine and Parasitology 79: 575­84. challenge facing the malaria epidemiologist today is to Brink, C. J. H. 1958. "Malaria Control in the Northern Transvaal." South describe malaria health outcomes in Africa and how these African Medical Journal 32: 800­08. Bruce-Chwatt, L. J. 1952. "Malaria in African Infants and Children in risks change from birth through adulthood according to Southern Nigeria." Annals of Tropical Medicine and Parasitology 46: the dependent factors of infection, immunity, and control. 173­200. Malaria | 209 Chandramohan, D., and B. M. Greenwood. 1998. "Is There an Interaction Government of the Colony and Protectorate of Kenya. 1935. Medical between Human Immunodeficiency Virus and Plasmodium falci- Department Annual Report, Including the Medical Research Laboratory parum?" International Journal of Epidemiology 27: 296­301. Annual Report for 1933. Nairobi, Kenya: Government Printers. Charlwood, J. D., M. Qassim, E. I. Elnsur, M. Donnelly, V. Petrarca, P. F. Government of Colony and Protectorate of Nigeria. 1934­35. Annual Billingsley, J. Pinto, and T. Smith. 2001. "The Impact of Indoor Reports of the Medical and Health Department for the Years 1932 and Residual Spraying with Malathion on Malaria in Refugee Camps in 1933. Lagos, Nigeria: Government Printers. Eastern Sudan." Acta Tropica 80: 1­8. Government of Colony and Protectorate of Sierra Leone. 1913­17. Annual Cockerell, O. C., A. L. Johnson, J. W. Sander, Y. M. Hart, D. M. Goodridge, Reports of the Medical Department for the Years ending 31st December and S. D. Shorvon. 1994. "Mortality from Epilepsy: Results from a 1912­1916. London: Waterlow & Sons. Prospective Population-Based Study." Lancet 344: 918­21. Government of Tanzania. 1997. Policy Implications of Adult Morbidity and Colbourne, M. J., and G. M. Edington. 1954. "Mortality from Malaria in Mortality: End of Phase 1 Report. Dar es Salaam, Tanzania: Ministry of Accra." Journal of Tropical Medicine and Hygiene 57: 203­10. Health. Colebunders, R., R. Ryder, H. Francis, W. Nekwei, Y. Bahwe, I. Lebughe, M. Granja, A. C., F. Machungo, A. Gomes, S. Bergstrom, and B. Brabin. 1998. Ndilu, G. Vercauteren, K. Nseka, and J. Perriens. 1991. "Seroconversion "Malaria-Related Maternal Mortality in Urban Mozambique." Annals Rate, Mortality and Clinical Manifestations Associated with the of Tropical Medicine and Parasitology 92: 257­63. Receipt of a Human Immunodeficiency Virus-Infected Transfusion in Greenberg, A. E., P. Nguyen-Dinh, J. M. Mann, N. Kabote, R. L. Kinshasa, Zaire." Journal of Infectious Diseases 164: 450­56. Colebunders, H. Francis, T. C. Quinn, et al. 1988. "The Association Coleman, R., L. Loppy, and G. Walraven. 2002. "The Treatment Gap and between Malaria, Blood Transfusions, and HIV Seropositivity in a Primary Health Care for People with Epilepsy in Rural Gambia." Pediatric Population in Kinshasa, Zaire." Journal of the American Bulletin of the World Health Organization 80: 378­83. Medical Association 259: 545­49. Craig, M. H., R. W. Snow, and D. le Sueur. 1999. "A Climate-Based Greenberg, A. E., M. Ntumbanzondo, N. Ntula, L. Mawa, J. Howell, and F. Distribution Model of Malaria Transmission in Sub-Saharan Africa." Davachi. 1989. "Hospital-Based Surveillance of Malaria-Related Parasitology Today 15: 105­11. Pediatric Morbidity and Mortality in Kinshasa, Zaire." Bulletin of the D'Alessandro, U., B. O. Olaleye, W. McGuire, P. Langerock, S. Bennett, World Health Organization 67: 189­96. M. K. Aikins, M. C. Thompson, M. K. Cham, B. A. Cham, and B. M. Greenwood, A. M., J. R. M. Armstrong, P. Byass, R. W. Snow, and B. M. Greenwood. 1995."Mortality and Morbidity from Malaria in Gambian Greenwood. 1992. "Malaria Chemoprophylaxis, Birth Weight and Children after the Introduction of an Impregnated Bed Net Child Survival." Transactions of the Royal Society of Tropical Medicine Programme." Lancet 345: 479­83. and Hygiene 86: 483­85. Deichmann, U. 1996. African Population Database. Digital database and Greenwood, B. M. 1990. "Populations at Risk." Parasitology Today 6: 188. documentation. National Center for Geographic Information and Analysis, Santa Barbara, CA. http://grid2.cr.usgs.gov/globalpop/africa/. Guyatt, H. L., and R. W. Snow. 2001a. "Malaria in Pregnancy as an Indirect Cause of Infant Mortality in Sub-Saharan Africa." Transactions of the Delacollette, C., and M. Barutwanayo. 1993. "Mortalité et morbidité aux Royal Society of Tropical Medicine and Hygiene 95: 569­76. jeunes ages dans une région a paludisme hyperendemique stable, com- mune de Nyanza Lac, Imbo Sud, Burundi." Bulletin de la Société de ------. 2001b. "The Epidemiology and Burden of Plasmodium falci- Pathologie Exotique 86: 1­7. parum­Related Anemia among Pregnant Women in Sub-Saharan Africa." American Journal of Tropical Medicine and Hygiene 64: Diseko, R. 1994. "Malaria in Botswana." Minutes of Annual Malaria 36­44. Conference, St. Lucia, South Africa, October. EANMAT (East African Network for Monitoring Antimalarial Therapy). Hauser, W. A., J. F. Annegers, and L. R. Elveback. 1980. "Mortality in 2003. "The Efficacy of Antimalarial Monotherapies Sulphadoxine- Patients with Epilepsy." Epilepsia 21: 399­412. Pyrimethamine and Amodiaquine in East Africa: Implications for Sub- Hay, S. I., J. Cox, D. J. Rogers, S. E. Randolph, D. I. Stern, D. G. Shanks, M. Regional Policy." Tropical Medicine and International Health 8: 860­67. F. Myers, and R. W. Snow. 2002. "Climate Change and the Resurgence English, M., F. Esamai, A. Wasunna, F. Were, A. Wamae, R. W. Snow, and N. of Malaria in the East African Highlands." Nature 415: 905­9. Peshu. 2004. "In-Patient Paediatric Care at the First Referral Level in Hay, S. I., C. A. Guerra, A. J. Tatem, P. M. Atkinson, and R. W. Snow. 2005. Kenya." Lancet 363: 1948­53. "Urbanization, Malaria Transmission and Disease Burden in Africa." Ettling, M., D. A. McFarland, L. J. Schultz, and L. Chitsulo. 1994. Nature Microbiology Review 3: 81­90. "Economic Impact of Malaria in Malawian Households." Tropical Hay, S. I., D. J. Rogers, J. F. Toomer, and R. W. Snow. 2000. "Annual Medicine and Parasitology 45: 74­79. Plasmodium falciparum Entomological Inoculation Rates (EIR) across Francesconi, P., M. Fabiani, M. G. Dente, M. Lukwiya, R. Okwey, J. Ouma, Africa: Literature Survey, Internet Access and Review." Transactions of R. Ochakachon, F. Cian, and S. Declich. 2001. "HIV, Malaria Parasites, the Royal Society of Tropical Medicine and Hygiene 94: 113­27. and Acute Febrile Episodes in Ugandan Adults: A Case-Control Study." Hill, A. V. S. 1992. "Malaria Resistance Genes: A Natural Selection." AIDS 15: 2445­50. Transactions of the Royal Society of Tropical Medicine and Hygiene 86: French, N., J. Nakiyingi, E. Lugada, C. Watera, J. A. Whitworth, and C. F. 225­26. Gilks. 2001. "Increasing Rates of Malarial Fever with Deteriorating Hoffman, I. F., C. S. Jere, T. E. Taylor, P. Munthali, J. R. Dyer, J. J. Wirima, Immune Status in HIV-1-Infected Ugandan Adults." AIDS 15: 899­906. S. J. Rogerson, et al. 1999. "The Effect of Plasmodium falciparum Gallup, J., and J. Sachs. 2001. "The Economic Burden of Malaria." Malaria on HIV-1 RNA Blood Plasma Concentration." AIDS 13: American Journal of Tropical Medicine and Hygiene 64S: 85­96. 487­94. Ghana VAST (Vitamin A Supplementation Treatment) Study Team. 1993. Holding, P. A., and R. W. Snow. 2001. "The Impact of Plasmodium falci- "Vitamin A Supplementation in Northern Ghana: Effects on Clinic parum Malaria on Performance and Learning: A Review of the Attendance, Hospital Admissions and Child Mortality." Lancet 342: Evidence." American Journal of Tropical Medicine and Hygiene 64 7­12. (Suppl. 1): 68­75. Government of Colony of Gold Coast. 1912­48. Annual Reports of the Holding, P. A., P. K. K. Wekulo. 2004. "Child Development: What Should Medical Department for the Years Ending 31st December 1912­1916. We Be Measuring and How Can We Measure It?" American Journal of London: Waterlow & Sons. Tropical Medicine and Hygiene 71 (Suppl. 2): 71­79. 210 | Robert W. Snow and Judy A. Omumbo INDEPTH Network 2002. Population and Health in Developing Countries. Moore, A., G. Herrera, J. Nyamongo, E. Lackritz, B. Nahlen, A. Oloo, Vol. 1: Population, Health and Survival at INDEPTH Sites. Ottawa: G. Opondo, R. Muga, and R. Janssen. 2001. "Estimated Risk of HIV International Development Research Centre. Transmission by Blood Transfusion in Kenya." Lancet 358: 657­60. Jaffar, S., A. Leach, A. M. Greenwood, A. Jepson, O. Muller, M. O. C. Ota, Mugisha, F., and J. Arinatwe. 2003. "Sleeping Arrangements and Mosquito K. Bojang, S. Obaro, and B. M. Greenwood. 1997. "Changes in the Net Use among Under Fives: Results from the Uganda Demographic Pattern of Infant and Childhood Mortality in Upper River Division, and Health Survey." Malaria Journal 2: 40­50. The Gambia, from 1989 to 1993." Tropical Medicine and International Mung'ala-Odera, V., R. W. Snow, and C. R. J. C. Newton. 2004. "The Health 2: 28­37. Burden of the Neuro-Cognitive Impairment Associated with Jillek-All, L., and H. T. Rwiza. 1992. "Prognosis of Epilepsy in a Rural Plasmodium Falciparum Malaria in Sub-Saharan Africa." American African Community: A 30-Year Follow-up of 164 Patients in an Journal of Tropical Medicine and Hygiene 71 (Suppl. 2): 55­63. Outpatient Clinic in Rural Tanzania. Epilepsia 33: 645­50. Murray, C. J. L., and A. D. Lopez. 1997. "Mortality by Cause for Eight Joy, D. A., X. R. Feng, J. B. Mu, T. Furuya, K. Chotivanich, A. U. Krettli, Regions of the World: Global Burden of Disease Study." Lancet 349: M. Ho, et al. 2003. "Early Origin and Recent Expansion of Plasmodium 1269­76. falciparum." Science 300: 318­21. Nabarro, D. N., and E. Taylor. 1998. "The `Roll Back Malaria' Campaign." Kahn, K., S. M. Tollman, M. Garenne, and J. S. S. Gear. 1999. "Who Dies Science 280: 2067­68. from What? Determining Cause of Death in South Africa's Rural Nevill, C. G., E. S. Some, V. O. Mung'ala, W. Mutemi, L. New, K. Marsh, North-East." Tropical Medicine and International Health 4: 433­41. C. Lengeler, and R. W. Snow. 1996. "Insecticide-Treated Bednets Kamwi, M. 1998. "Report on Malaria Situation." Ministry of Health and Reduce Mortality and Severe Morbidity from Malaria among Children Social Services Epidemiology Section, Government of Namibia. on the Kenyan Coast." Tropical Medicine and International Health 1: 139­46. Kleinschmidt, I., J. A. Omumbo, O. Briet, N. van de Giesen, N. Sogoba, N. K. Mensah, P. Windermeijer, M. Moussa, and T. Teuscher. 2001. "An Newton, C. R. J. C. and S. Krishna. 1998. "Severe Falciparum Malaria in Empirical Malaria Distribution Map for West Africa." Tropical Children: Current Understanding of Pathophysiology and Supportive Medicine and International Health 6: 779­86. Treatment." Pharmacology and Therapeutics 79: 1­53. Korenromp, E. L., J. R. M. Armstrong-Schellenberg, B. G. Williams, Nussenblatt, V., and R. D. Semba. 2002. "Micronutrient Malnutrition and B. Nahlen, and R. W. Snow. 2004. "Impact of Malaria Control on the Pathogenesis of Malarial Anemia." Acta Tropica 82: 321­37. Childhood Anemia in Africa--A Quantitative Review." Tropical Omumbo, J. A., C. A. Guerra, S. I. Hay, and R. W. Snow. 2004. "The Medicine and International Health 9: 1050­65. Relationship between the Parasite Ratio in Childhood and Climate Korenromp, E. L., B. G. Williams, E. Gouws, C. Dye, and R. W. Snow. 2003. Estimates of Malaria Transmission in Kenya." Malaria Journal 3: 1­8. "Measuring Trends in Childhood Malaria Mortality in Africa: A New Parise, M. E., J. G. Ayisi, B. L. Nahlen, L. J. Schultz, J. M. Roberts, A. Misore, Assessment of Progress toward Targets Based on Verbal Autopsy." R. Muga, A. J. Oloo, and R. W. Steketee. 1998. "Efficacy of Sulfadoxine- Lancet Infectious Diseases 3: 349­58. Pyrimethamine for Prevention of Placental Malaria in an Area of Lackritz, E. M., C. C. Campbell, T. K. Ruebush II, A. W. Hightower, Kenya with a High Prevalence of Malaria and Human Immunodefi- W. Wakube, R. W. Steketee, and J. B. O. Were. 1992. "Effect of Blood ciency Virus Infection." American Journal of Tropical Medicine and Transfusion on Survival among Children in a Kenyan Hospital." Lancet Hygiene 59: 813­22. 340: 524­28. Pasha, O., J. D. Rosso, M. Mukaka, and D. Marsh. 2003. "The Effect of Le Sueur, D., B. L. Sharp, and C. C. Appleton. 1993. "Historical Perspective Providing Fansidar (sulfadoxine-pyrimethamine) in Schools on of the Malaria Problem in Natal with Emphasis on the Period Mortality in School-Age Children in Malawi." Lancet 361: 577­8. 1928­1932." South African Journal of Science 89: 232­39. Pelletier, D. L., E. A. Frongillo, and J. P. Habicht. 1993. "Epidemiologic LSDI (Lubombo Special Development Initiative). 2001. "Annual Report, Evidence for a Potentiating Effect of Malnutrition on Child Mortality." 2001." LSDI Malaria Control. Medical Research Council Report, American Journal of Public Health 83: 1130­33. Durban, South Africa. Pelletier, D. L., E. A. Frongillo, D. G. Schroeder, and J. P. Habicht. 1995. "The Effects of Malnutrition on Child Mortality in Developing Luxemburger, C., R. McGready, A. Kham, L. Morison, T. Cho, T. Countries." Bulletin of the World Health Organization 73: 443­48. Congsuphajaisiddhi, N. J. White, and F. Nosten. 2001. "Effects of Malaria during Pregnancy on Infant Mortality in an Area of Low Philips-Howard, P., and A. B. Bjorkman. 1990. "Ascertainment of Risk of Malaria Transmission." American Journal of Epidemiology 154: 459­65. Serious Adverse Reactions Associated with Chemoprophylactic Antimalarial Drugs." Bulletin of World Health Organization 68: Malaney, P., A. Speilman, and J. Sachs. 2004. "The Malaria Gap." American 493­504. Journal of Tropical Medicine and Hygiene 71 (Suppl. 2): 141­6. Philips-Howard, P., and L. J. West. 1990. "Serious Adverse Drug Reactions Marsh, K., D. Forster, C. Waruiru, I. Mwangi, M. Winstanley, V. Marsh, to Pyrimethamine-Sulphadoxine, Pyrimethamine-Dapsone and to C. Newton, et al. 1995. "Life-Threatening Malaria in African Children: Amodiaquine in Britain." Journal of the Royal Society of Medicine 83: Clinical Spectrum and Simplified Prognostic Criteria." New England 82­85. Journal of Medicine 332: 1399­404. Piotti, A. 1997. "Memo on Collection of Malaria Clinical Case Incidence Marsh, K., and R. W. Snow. 1999. "Malaria Transmission and Morbidity." Rates, December 1995 to April 1996 and December 1996 to April Parassitologia (Rome) 41: 241­46. 1997." Report, Ministry for Health, Zimbabwe, July 4. McCarthy, D. H. W., and Y. Wu. 2000. Malaria and Growth. Policy Research Premji, Z., P. Ndayanga, C. Shiff, J. Minjas, P. Lubega, and J. MacLeod. Working Paper 2303, World Bank, Washington, DC. 1997. "Community Based Studies on Childhood Mortality in a Malaria Menendez, C. 1995."Malaria during Pregnancy: A Priority Area of Malaria Holoendemic Area on the Tanzanian Coast." Acta Tropica 63: 101­9. Research and Control." Parasitology Today 11: 178­83. Reynolds, J. E. F., ed. 1993. Martindale: The Extra Pharmacopoeia. 13th ed. Menendez, C., A. F. Fleming, and P. L. Alonso. 2000. "Malaria-Related London: Pharmaceutical Press. Anaemia." Parasitology Today 16: 469­76. Robert, V., K. MacIntyre, S. Keating, J. F. Trape, J. B. Duchemin, Molineaux, L. 1997. "Malaria and Mortality: Some Epidemiological M. Wilson, and J. C. Beier. 2003. "Malaria Transmission in Urban Considerations." Annals of Tropical Medicine and Parasitology 91: Sub-Saharan Africa." American Journal of Tropical Medicine and 811­25. Hygiene 68: 169­76. Malaria | 211 Salum, F. M., T. J. Wilkes, K. Kivumbi, and C. F. Curtis. 1994. "Mortality of ------. 1997b. "The Effects of Malaria Control on Nutritional Status in Under Fives in a Rural Area of Holoendemic Malaria Transmission." Infancy." Acta Tropica 65: 1­10. Acta Tropica 58: 29­34. Snow, R. W., J. F. Trape, and K. Marsh. 2001. "The Past, Present and Future Savarit, D., K. M. De Cock, R. Schutz, S. Konate, E. Lackritz, and A. of Childhood Malaria Mortality in Africa." Trends in Parasitology 17: Bondurand. 1992."Risk of HIV Infection from Transfusion with Blood 593­97. Negative for HIV Antibody in a West African City." British Medical Snow, R. W., R. E. M. Williams, J. E. Rogers, V. O. Mung'ala, and N. Peshu. Journal 305: 498­501. 1994. "The Prevalence of Epilepsy among a Rural Kenyan Population: Schellenberg, J. R. M. A., C. G. Victora, A. Mushi, D. De Savigny, D. Its Association with Premature Mortality." Tropical and Geographical Schellenberg, H. Mshinda, and J. Bryce. 2003. "Inequities among the Medicine 46: 175­79. Very Poor: Health Care Children in Rural Southern Tanzania." Lancet Steketee, R. W., B. L. Nahlen, M. E. Parise, and C. Menendez. 2001. 361: 561­66. "The Burden of Malaria in Pregnancy in Malaria-Endemic Areas." Schwartlander, B. 1997. "Global Burden of Disease." Lancet 350: 141­42. American Journal of Tropical Medicine and Hygiene 64 (Suppl.): Shanks, G. D., K. Biomondo, S. I. Hay, and R. W. Snow. 2000. "Changing 28­35. Patterns of Clinical Malaria Since 1965 among a Tea Estate Population Steketee, R. W., J. J. Wirima, P. B. Bloland, B. Chilima, J. H. Mermin, Located in the Kenyan Highlands." Transactions of the Royal Society of L. Chitsulo, and J. G. Breman. 1996. "Impairment of a Tropical Medicine and Hygiene 94: 253­55. Pregnant Woman's Acquired Ability to Limit Plasmodium Shepard, D. S., M. B. Ettling, U. Brinkmann, and R. Sauerborn. 1991. "The falciparum by Infection with Human Immunodeficiency Virus Economic Cost of Malaria in Africa." Tropical Medicine and Type-1." American Journal of Tropical Medicine and Hygiene 55 Parasitology 42: 199­203. (Suppl.): 42­49. Shulman, C. E. 1999. "Malaria in Pregnancy: Its Relevance to Safe- Steketee, R. W., J. J. Wirima, A. W. Hightower, L. Slutsker, D. L. Heymann, Motherhood Programmes." Annals of Tropical Medicine and and J. G. Breman. 1996. "The Effect of Malaria and Malaria Prevention Parasitology 93 (Suppl. 1): S59­66. in Pregnancy on Offspring Birthweight, Prematurity, and Intrauterine Growth Retardation in Rural Malawi." American Journal of Tropical Smith, P. G., and R. H. Morrow. 1996. Field Trials of Health Interventions in Medicine and Hygiene 55 (Suppl.): 33­41. Developing Countries: A Toolbox. 2nd ed. London: Macmillan Education. Sturchler, D. 1989. "How Much Malaria is There World-Wide?" Snow, R. W., M. H. Craig, U. Deichmann, and K. Marsh. 1999. "Estimating Parasitology Today 5: 39­40. Mortality, Morbidity and Disability Due to Malaria among Africa's Talisuna, A. O., P. Bloland, and U. D'Alessandro. 2004. "History, Dynamics Non-Pregnant Population." Bulletin of the World Health Organization and Public Health Importance of Malaria Parasite Resistance." Clinical 77: 624­40. Microbiology Reviews 17: 235­54. Snow, R. W., M. H. Craig, C. R. J. C. Newton, and R. W. Steketee. 2003."The Ter Kuile, F. O., D. J. Terlouw, S. K. Kariuki, P. A. Phillips-Howard, L. B. Public Health Burden of Plasmodium falciparum Malaria in Africa: Mirel, W. A. Hawley, J. F. Friedman, et al. 2003. "Impact of Permethrin- Deriving the Numbers." Disease Control Priorities Project Working Treated Bednets on Malaria, Anemia and Growth in Infants in an Area Paper 11, Fogarty International Center, National Institutes of Health, of Intense Perennial Malaria Transmission in Western Kenya." Bethseda, Md. http://www.fic.nih.gov/dcpp. American Journal of Tropical Medicine and Hygiene 68 (Suppl. 4): Snow, R. W., and H. M. Gilles. 2002. "The Epidemiology of Malaria." In 68­77. Bruce-Chwatt's Essential Malariology, 4th ed., ed. D. A. Warrell and Trape, J. F., E. Lefebvre-Zante, F. Legros, G. Ndiaye, H. Bouganali, H. M. Gilles, 85­106. London: Arnold Publishers. P. Druille, and G. Salem. 1992. "Vector Density Gradients and the Snow, R. W., E. Gouws, J. Omumbo, B. Rapuoda, M. H. Craig, F. C. Tanser, Epidemiology of Urban Malaria in Dakar, Senegal." American Journal D. le Sueur, and J. Ouma. 1998. "Models to Predict the Intensity of of Tropical Medicine and Hygiene 47: 181­89. Plasmodium falciparum Transmission: Applications to the Burden of Trape, J. F., G. Pison, M. P. Preziosi, C. Enel, A. Desgrees du Lou, V. Del Disease in Kenya." Transactions of the Royal Society of Tropical Medicine Aunay, B. Samb, E. Lagarde, J. F. Molez, and F. Simondon. 1998. and Hygiene 92: 601­6. "Impact of Chloroquine Resistance on Malaria Mortality." Comptes Snow, R. W., E. L. Korenromp, and E. Gouws. 2004. "Pediatric Mortality in Rendus de l'Académie des Sciences Paris, Série III 321: 689­97. Africa: Plasmodium Falciparum Malaria as a Cause or a Risk?" Tsoka, J. M., B. L. Sharp, and I. Kleinschmidt. 2002. "Malaria Mortality in American Journal of Tropical Medicine and Hygiene 71 (Suppl. 2): a High Risk Area of South Africa." Paper presented at the Third MIM 16­24. Pan-African Malaria Conference: "Global Advances in Malaria Snow, R. W., and K. Marsh. 2002. "The Consequences of Reducing Research: Evidence-Based Decision Making for Malaria Control Plasmodium falciparum Transmission in Africa." Advances in Policy," Arusha, Tanzania, November. Parasitology 52: 235­64. United Nations Population Division. 2001. World Population Prospects: Snow, R. W., C. S. Molyneux, P. A. Warn, J. Omumbo, C. G. Nevill, S. The 2000 Revision and World Urbanization Prospects: The 2001 Gupta, and K. Marsh. 1996. "Infant Parasite Rates and Revision. http://esa.un.org/unpp. Immunoglobulin M Seroprevalence as a Measure of Exposure to van Eijk, A. M., J. G. Ayisi, and F. O. ter Kuile. 2001. "Human Plasmodium falciparum during a Randomized Controlled Trial of Immunodeficiency Virus Increases the Risk of Malaria in Women of Insecticide-Treated Bed Nets on the Kenyan Coast." American Journal All Gravidities in Kisumu, Kenya." Abstract 405 in Proceedings of 50th of Tropical Medicine and Hygiene 55: 144­49. Annual Meeting of the American Society of Tropical Medicine and Snow, R. W., V. O. Mung'ala, D. Forster, and K. Marsh. 1994. "The Role of Hygiene, Atlanta, Ga. the District Hospital in Child Survival at the Kenyan Coast." African van Eijk, A. M., J. G. Ayisi, F. O. ter Kuile, A. Misore, J. A. Otieno, M. S. Journal of Health Sciences 1: 71­75. Kolczak, P. A. Kager, R. W. Steketee, and B. L. Nahlen. 2001. "Human Snow, R. W., J. A. Omumbo, B. Lowe, S. M. Molyneux, J. O. Obiero, A. Immunodeficiency Virus Seropositivity and Malaria as Risk Factors for Palmer, M. W. Weber, et al. 1997a. "Relation between Severe Malaria Third-Trimester Anemia in Asymptomatic Pregnant Women in Morbidity in Children and Level of Plasmodium falciparum Western Kenya." American Journal of Tropical Medicine and Hygiene 65: Transmission in Africa." Lancet 349: 1650­54. 623­30. 212 | Robert W. Snow and Judy A. Omumbo Verhoeff, F. H., B. J. Brabin, C. A. Hart, L. Chimsuku, P. Kazembe, and R. and Clinical Episodes in Adults in Rural Uganda: A Cohort Study." L. Broadhead. 1999. "Increased Prevalence of Malaria in HIV-Infected Lancet 356: 1051­56. Pregnant Women and Its Implications for Malaria Control." Tropical Xiao, L., S. M. Owen, D. L. Rudolph, R. B. Lal, and A. A. Lal. 1998. Medicine and International Health 4: 5­12. "Plasmodium falciparum Antigen-Induced Human Immunodeficiency Warrell, D. A. 2002. "Clinical Features of Malaria." In Bruce-Chwatt's Virus Type 1 Replication Is Mediated through Induction of Essential Malariology, 4th ed., ed. D. A. Warrell and H. M. Gilles, Tumor Necrosis Factor-Alpha." Journal of Infectious Diseases 177: 191­205. London: Arnold Publishers. 437­45. Whitworth, J., D. Morgan, M. Quigley, A. Smith, B. Mayanja, H. Eotu, N. Zielinski, J. J. 1974. "Epilepsy and Mortality Rate and Cause of Death." Omoding, M. Okongo, S. Malamba, and A. Ojwiya. 2000. "Effect of Epilepsia 15: 191­201. HIV-1 and Increasing Immunosuppression on Malaria Parasitaemia Malaria | 213 Chapter 15 Onchocerciasis Uche Amazigo, Mounkaila Noma, Jesse Bump, Bruce Benton, Bernhard Liese, Laurent Yaméogo, Honorat Zouré, and Azodoga Seketeli Onchocerciasis, commonly called river blindness, is a para- There are many vectors in the Western hemisphere, but sitic disease particularly prevalent in Africa, where more S. ochraceum, S. metallicum, S. oyapockense, S. guianense, than 99 percent of all cases occur. In total, 30 countries are and S. exiguum are among the most important. In the island infested, ranging from Senegal across to Ethiopia in the of Bioko, the vector is S. yahense, Bioko form. north and as far south as Angola and Malawi. Blackflies serve as the intermediate host of the parasite. Flies ingest immature worms when they bite infected people. As the worms live in the fly, they mature sexually TRANSMISSION over the course of a week. Then, should the fly bite a person, the maturing worm will grow to adulthood inside the Onchocerciasis is caused by worms, Onchocerca volvulus. human body. Upon finding mating partners, the adults The adult worms measure nearly a meter long and live in become encapsulated and produce more immature worms, coiled mating pairs in nodules under the skin. Reproducing completing the full transmission cycle. adult females spawn about 2,000 immature worms every day. These tiny juvenile worms migrate throughout the skin and eyes, causing the various symptoms of the disease. THE DISEASE BURDEN Although they are damaging, these immature worms cannot mature to adulthood without being transmitted by a black- The most severe consequences of onchocerciasis is blind- fly of the genus Simulium. This fly breeds in rapidly flowing ness, which may affect one-third of the adult population of streams and rivers, and thus the name "river blindness." The the most highly affected communities. The prevention of most important vector is Simulium damnosum sensus lato, blindness was the main clinical reason for initiating the which has a wide range throughout Africa and the Middle Onchocerciasis Control Programme (OCP) in West Africa East. In East Africa, S. neavei also transmits onchocerciasis. in 1974 (Benton et al. 2002). Onchocerciasis causes several 215 symptoms, including unrelenting itching, physical scars offered some protection from further infection, but then from constant scratching, de-pigmentation and thickening farmers struggled with poor soil and water shortages on of the skin, impaired vision, and complete blindness. overcrowded lands. Onchocerciasis often ultimately sent a Important pioneering studies in the last decade, spon- prosperous community into poverty. Armed with this new sored by the United Nations Development Programme knowledge about the economic impact, and with many (UNDP), the World Bank, and the World Health fertile lands and water sources off-limits because of endemic Organization (WHO) Special Programme for Research and onchocerciasis, development agencies made the disease a Training in Tropical Diseases (TDR), have shown that priority. onchocercal skin disease (OSD) is associated with a greater degree of morbidity than was hitherto appreciated. These studies demonstrated that severe OSD causes suffering to RAPID EPIDEMIOLOGICAL MAPPING millions of people, particularly to those in the forest zone, OF ONCHOCERCIASIS where the blinding form of the disease is less prevalent (Amazigo 1993; WHO 1995b). OSD not only causes psy- Continued research has led to the development of Rapid chosocial problems, ostracism, and stigma (Brieger et al. Epidemiological Mapping of Onchocerciasis (REMO), a 1998b; Okello, Ovuga, and Ogwal Okeng 1995; Ovuga et al. tool that provides data on the distribution and prevalence of 1995) but also has a demonstrably negative socioeconomic the disease. REMO is a simple, noninvasive, and practicable impact on the productivity of farmers, breastfeeding, and process that is easy to apply over a wide range of bioecolog- school attendance (Amazigo 1994; Benton 1998; Kim et al. ical zones, with no sociocultural or religious restrictions. 1997; Oladepo et al. 1997; Vlassoff et al. 2000). This tool was vital to rapid mapping and accelerating the The relative contributions of both ocular and dermal scaling-up process of identifying endemic villages for mass symptoms to the burden of onchocercal disease and their treatment with ivermectin in control areas. Integration of socioeconomic consequences are substantial (Kale 1998). REMO data into the Geographical Information System The unbearable itching and blindness that accompany the (GIS) allowed delineation of zones of different levels of disease hinder the ability of individuals to contribute to endemicity, which is an important step in the planning their own well-being, and they undermine the emotional process for onchocerciasis control. Zones, and communities and economic health of the household and the commu- in these zones, are included or excluded from the community- nity. Consequently, onchocerciasis, which predominantly directed treatment with ivermectin (CDTI), depending on affects poor people in remote areas, can be directly linked whether their levels of onchocercal endemicity reach the to poverty. threshold set by the control program. Prevalence rates of infection in the OCP areas prior to control activities were high in all the countries. THE SOCIOECONOMIC TOLL In the 1970s, when investigations of onchocerciasis in the POPULATION AT RISK AND POPULATION endemic villages and districts of West Africa began, scien- INFECTED tists made astonishing and disturbing discoveries. More than 60 percent of the savanna population carried the para- In 1995 it was estimated that 18 million persons were in- site; 10 percent of the adult population and half of the males fected, of whom 270,000 were blind and about 500,000 per- over 40 years of age were blind, 30 percent of the people sons severely visually impaired (WHO 1995a). However, the were visually impaired, and early signs of onchocerciasis large-scale REMO that was undertaken by the African were common among children (WHO 1973, 1987, 1995a). Programme for Onchocerciasis Control (APOC), covering Scientists revealed the huge socioeconomic consequences the 19 remaining endemic countries in Africa, provides of the high infection rates they had found.As village blindness recent and more accurate estimates. From the REMO results reached epidemic proportions, it left too few able-bodied it was estimated in 2004 that about 87 million persons were people to tend fields. Food shortages and economic collapse at high risk of contracting onchocerciasis in APOC coun- forced residents to abandon homelands in fertile river val- tries. This figure is higher than the number at risk (about leys. Moving to hardscrabble highlands and forested areas 58 million persons) reported two years earlier by Noma and 216 | Uche Amazigo, Mounkaila Noma, Jesse Bump, Bruce Benton, Bernhard Liese, Laurent Yaméogo, Honorat Zouré, and Azodoga Seketeli colleagues (2002), when countrywide mapping of the dis- yielded enormous advantages, including rapid expansion of ease had not been carried out in Angola and Burundi and ivermectin distribution to remote communities. The part- had been implemented (only partially) in very few areas in ners have ably demonstrated how to deliver medicines to the Democratic Republic of Congo and Sudan. reach those who need them most, the poorest. With lessons The results of REMO, also used to determine the number from the OCP, the partners of APOC are making efforts to of infected persons in the APOC countries, showed that mentor national or local nongovernmental organizations the mean infection rate was 38.2 percent, or approximately (NGOs) in different countries to play major roles in sus- 37 million people. This represents almost three times the taining ivermectin distribution by supporting the health number estimated (13.7 million people) at the planning system, especially after the withdrawal of external support. phase of the APOC program, when only partial data were available for few countries. Vector Control Onchocerciasis control is complicated by the long life of the ONCHOCERCIASIS CONTROL PROGRAMS adult worms, which remain fertile throughout most of their 10- to 15-year lifespan. The immature worms live in the skin In 1970, funded by the UNDP, a team of WHO scientists for about two years, but their numbers are continually and consultants began to lay the technical groundwork for a refreshed as long as adults are alive in the body. This means major regional initiative to defeat river blindness. By 1972 that even with instant and complete transmission control, the international development community mobilized funds the disease would not die out naturally for 15 years; hence, to aid in the control of the disease. In 1974 the affected control attempts must last at least the life span of the adult countries and four agencies (the World Bank, the WHO, worm. The control program in West Africa, the OCP, initially the Food and Agriculture Organization [FAO], and the attacked the disease by killing the larvae of the flies that UNDP) launched an unprecedented partnership. The "river transmit the worms. It depended on killing these larvae over blindness partnership" has had two distinct phases: the OCP a long enough time that the adult worms would all die out. between 1974 and 2002 and APOC between 1996 and 2010. Then, when fly control stopped, biting flies would no longer The OCP had a dual mandate: to eliminate river blindness ingest any parasites, and the transmission cycle would be as a public heath problem and as an obstacle to socioeco- broken. The key to this approach lies in reducing the fly nomic development. The APOC mandate in East Africa and population sufficiently to stop transmission and then central Africa does not express the same socioeconomic sustaining the effort for two decades or more. development needs as did the OCP mandate in West Africa. The principal strategy of the OCP from its inception was This is because the strain of parasite prevalent outside the vector control. Covering Burkina Faso and six neighboring savanna belt is less likely to cause blindness. Instead, it exacts countries, the control began in 1974 in West Africa as a large its toll in greater skin disease. The stigma and disability from regional project. Vector control, treating the breeding sites of these dermatologic effects are difficult to quantify, and disease-transmitting flies with larvicides, was the only avail- humanitarian reasons alone were more than sufficient to able approach. justify the expense of control. Prior control attempts dating back to the 1950s had shown that the disease is transmitted on a regional scale. The first projects were small, and the savanna was consis- Partnership of the Control Programs tently reinfested. Accordingly, the OCP was established as a The onchocerciasis control programs in Sub-Saharan large program to cover entire endemic zones. Even this Africa--the OCP and, in particular, APOC--developed a ambitious start was not sufficient; in 1986 the program dou- unique partnership structure. More than 80 partners are bled in size and was expanded to cover 11 countries in all. involved in this rich coalition, including 26 donors; Since blackflies migrate across international borders, the 30 African countries; a major pharmaceutical firm, Merck & affected governments and international experts were con- Co.; and 12 major nongovernmental development organiza- vinced that only a regional program could control river tions (NGDOs) over 120,000 local communities within the blindness. Thus, the OCP targeted seven West African coun- OCP and APOC ambit of operations. This broad coalition is tries (Benin, Burkina Faso, Côte d'Ivoire, Ghana, Mali, complex to maintain but has created synergies that have Niger, and Togo; figure 15.1). With the collaboration and Onchocerciasis | 217 Figure 15.1 Prevalence of Onchocerciasis Infection in 1974 The favorable results of these studies led to the OCP's and Gradual Expansion, 1977­92 adopting mass distribution of Mectizan as an adjunct to vector control. In 1990 the program began full-scale distri- bution in extension areas--to the south and west of the original core area in West Africa--using mobile teams in jeeps plus local health staff support. In this first step toward scaling up, OCP-paid local health professionals called communities to a central location for dosing. In more than 30 river basins, therapeutic coverage averaged about 65 percent in 1987, improving to more than 70 percent by 1995. However, it was expensive to use trained health staff at the local level. In light of high, recurring costs, the program considered various cost-recovery schemes to no avail. The answer to the high cost of mobile teams arrived 5% 15% 30% 45% 60% indirectly. Invariably, when drugs were distributed, some villagers were away, either hunting, working, or traveling. In Source: Onchocerciasis Coordination Unit, World Bank. Note: The dotted green line indicates the original seven-country control area. The solid green response, the program authorized the mobile teams to leave line encloses 11 countries and shows the extensions made during program implementation. doses upon departure for absent community members, once it became clear that ivermectin's safety profile allowed unsu- pervised dosing. In the second step toward scaling up, political commitment of these nations, control operations national health services combined with local health staff to were planned. As the primary method of control, aircraft distribute the drug, forming a community-based distribu- would spray environmentally safe larvicides around fast- tion approach. flowing rivers--the breeding grounds of the intermediate In some areas the OCP used ivermectin alone (Awadzi host of the disease, the blackfly. et al. 1985; Dadzie et al. 1990, 1991; Remme et al. 1989, 1990; Initially, operations covered 660,000 square kilometers Whitworth et al. 1991,1992) and effectively so. Ivermectin is in seven countries, an area believed to be large enough to effective against only the juvenile parasites, killing 95 per- contain the blackfly vector (WHO 1995a). However, in cent with one dose. Even when a patient takes ivermectin, 1975, after three initial months of successful operations, the adult worms continue to live, churning out offspring. many migrant blackflies from untreated watercourses reap- However, because the juvenile parasites cause the disease, peared, threatening to reintroduce the disease into the pro- ivermectin relieves the symptoms and allows the body to gram area. Scientists found that the flies were coming from begin healing itself. Treatment with ivermectin is required up to 600 kilometers away from the area being treated only once per year but must be taken for as long as any (Le Berre et al. 1990; WHO 1995a). In response, the pro- adults are still alive, up to 15 years. By killing almost all the gram extended operations to another four West African immature worms, ivermectin also dramatically lowers the countries (Guinea, Guinea-Bissau, Senegal, and Sierra chance of parasite ingestion by biting flies. Leone; figure 15.1). The program area increased geographic The application of the two strategies by the OCP has led coverage from 780,000 square kilometers to 1.3 million to the virtual elimination of onchocerciasis as a public square kilometers, enabling the campaign to increase the health problem and as an obstacle to socioeconomic devel- number of people protected from 10 million to 30 million. opment in 10 of the 11 countries in which the program was carried out (figure 15.2), leading to the recognition of OCP as one of the most successful programs in the history of Ivermectin Treatment and Approaches to Disease Control development assistance (Kim and Benton 1995). Figure 15.3 The advent of ivermectin (Mectizan) and its donation by shows the other health interventions implemented using Merck & Co. in 1987, for as long as needed, provided a the CDTI network in APOC countries in 2004. By the end of second string to OCP's control operations. The first exten- the OCP, the program had covered 11 countries, protecting sive field studies on the suitability of the drug for use on a 40 million persons at risk and 1.3 million square kilometers mass scale were conducted by the OCP (Awadzi et al. 1985). of land. By the end of APOC in 2010, the program's two 218 | Uche Amazigo, Mounkaila Noma, Jesse Bump, Bruce Benton, Bernhard Liese, Laurent Yaméogo, Honorat Zouré, and Azodoga Seketeli Figure 15.2 Prevalence of Onchocerciasis Infection in 2002 together to form the NGDO Coordination Group for Ivermectin Distribution. By 1995 the group needed consid- erably more resources than they could generate on their own for a significant expansion in the scope of their activi- ties. Furthermore, the various NGDOs and programs used mobile teams and the clinic-based and community-based treatment with ivermectin in the distribution of the drug. These methods were proved inappropriate or not cost- effective for large-scale sustainable distribution of the drug for several years. In 1995 the Task Force on Onchocerciasis Operational Research of the UNDP, World Bank, and WHO Special Program for Research and Training in Tropical Diseases, in collaboration with the OCP, addressed the prob- 5% 15% 30% 45% 60% lem in a multicountry research study (WHO 1995b). The study concluded that community-directed treatment Source: Onchocerciasis Coordination Unit, World Bank. Note: The dotted green line indicates the original seven-country control area. The solid green with ivermectin (CDTI) was feasible and effective in a wide line encloses 11 countries. The map shows the larger area covered in 2002, and the much lower range of geographical and cultural settings in Africa, and rate of infection obtained after 28 years of OCP, from 1974 to 2002. likely to be replicable in other endemic communities in Africa. It recommended that this approach become a princi- Figure 15.3 Scaling Up with Additional Interventions pal method for onchocerciasis control in Africa. (WHO 1996). The scaling up to community-directed treatment A community-directed network offers a key entry point for many health interventions in the most remote, rural communities. Phase II reaches the poorest of the poor in areas where National was a major turning point for the onchocerciasis control Health Services are weak or nonexistent. programs. Two successful regional programs, APOC and the Lymphatic filariasis treatment Onchocerciasis Elimination Program for the Americas Vitamin A distribution (OEPA), were introduced to distribute ivermectin (Remme Schistosomiasis treatment 1995; Richards et al. 2001). APOC was launched in 1995 to Guinea worm intervention cover 19 countries in the remainder of infested Africa. In Immunizations (polio, measles, others) 1997 the program adopted the CDTI as its principal strate- Eye care (cataract identification, primary eye care) gy for ivermectin distribution. In 2001, 42 million tablets Malaria bednet distribution of ivermectin donated by Merck & Co. were used in the HIV/AIDS and reproductive health 28 countries in Sub-Saharan Africa for onchocerciasis con- trol. Ivermectin is distributed by community workers, trained and supported by the external partners. Source: Onchocerciasis Coordination Unit, World Bank. In 1996 APOC had approved four CDTI projects, and by the end of 2003 the Programme totaled 107 projects to be phases will have protected an estimated 150 million people implemented (table 15.1), of which 97 were CDTI projects. in 30 countries. Sixty-two projects delivered more than 32 million treat- The initial efforts at mass distribution of ivermectin out- ments in 2003 alone (table 15.1). By 2010 the total treat- side OCP were made by NGDOs. The first of the NGDO- ments are expected to reach 90 million people. The distri- facilitated ivermectin distribution programs in Africa was bution network is also being tested to deliver other health established in Nigeria in 1989 (Duke and Dadzie 1993). Many interventions. This enticing possibility opens the door to of the pioneering NGDOs were already well known through further scaling up to help control other diseases and pres- their activities in prevention of blindness and were already ents the opportunity to deliver other basic health interven- with the WHO Prevention of Blindness Programme. tions in the river-blindness areas, all of which are remote, The NGDOs soon recognized the need to coordinate rural, and poor. Most are not reached by other programs, their separate and independent efforts if they were to and some are not reached by the national governments achieve their common goal. Therefore, in 1992 they came themselves. Onchocerciasis | 219 Table 15.1 Scaling up APOC, 1996­2003 1996 1997 1998 1999 2000 2001 2002 2003 Total Projects approved (including CDTI, 4 25 16 12 6 6 11 27 107 Vector Elimination, and National Onchocerciasis Task Forces HQ support) Annual treatments n.a. 1.5 14.1 16.9 20.4 24.6 29.0 32.2 138.7 Geographic coverage (percent)a n.a. 36.6 39.5 42.3 73.0 69.8 75.3 84.8 n.a. Therapeutic coverage (percent)b n.a. 52.7 54.2 57.8 58.6 54.8 58.2 68.1 n.a. Source: APOC. Note: n.a. not applicable. a. Geographic coverage is the percentage of communities treated per total communities at high risk. b. Therapeutic coverage is the percentage of people covered per total population at high risk. In 2003, through community-directed treatment, APOC Benefits of the Onchocerciasis Control Programs countries achieved ivermectin therapeutic coverage of The 20 percent economic rate of return of OCP from the 68.1 percent on average in the targeted communities. It inception of phase 1 in 1975 to 2002 compares well with augmented these distribution efforts with a strong commit- other development projects, including those outside the ment to increasing capacity. health sector (Kim and Benton 1995). In the study by Kim The introduction of ivermectin presented challenges and and Benton (1995) the benefits taken into account were an opportunities that became a catalyst for scaling up on all increase in the labor force due to prevention of blindness levels. It transformed onchocerciasis control from a techno- (25 percent of benefits) and increased land use (75 percent logically driven categorical health initiative to a community- of benefits). directed process of treatment and empowerment of By 2010 the total budget for the OCP and APOC will communities. This grassroots approach contributed, as amount to approximately US$735 million in donor shown in table 15.1, to high coverage of the population and financing; these funds were allocated primarily to larvicid- empowered communities to take charge of their own health. ing and entomological evaluation. Other costs were for It also planted the seeds for sustainability--absolutely vital administration, including extensive meetings aimed at for a disease that must be treated for at least 15 years to ensuring transparency within the wide-ranging partnership; interrupt transmission. Systematic assessments of 48 CDTI ivermectin delivery; training; and research and develop- projects in 10 countries (Cameroon, Chad, the Democratic ment. The cost of protection from OCP operations per Republic of Congo, Ethiopia, Malawi, Nigeria, the Republic person per year was well under US$1. The target cost per of Congo, Sudan, Tanzania, and Uganda) at four levels-- treatment is about US$0.15 for APOC. namely, the central level (the state, regional, and provincial Normally, a 10 percent rate of return for World Bank levels of the different health systems in Sub-Saharan Africa), projects in the "productive sector" (excluding social proj- the district and local government authorities level, the ects, such as education and health) is considered a success. subdistrict and frontline health facility level, and the com- By 2010, the economic rate of return for APOC is expected munity level--were undertaken between 2002 and 2004. to reach 18 percent. Moreover, every U.S. dollar invested in These sustainability assessments were carried out with out- APOC activities is expected to add 27 labor-productive days side technical assistance. A total of 304 district and local between 1996 and 2017. government authorities and 468 communities were evalu- ated. The performance of health care service providers and community workers was assessed using quality-of- CONCLUSION implementation indicators designed for the multicountry evaluation of CDTI. The results of these in-depth evalua- Over the last 30 years a large international partnership has tions show that 35 (73 percent) of the 48 projects are successfully attacked onchocerciasis. This partnership has making satisfactory progress toward sustainability, and 13 defeated the disease in 10 of the 11 countries in West Africa (27 percent) are not. and is making progress in the remaining endemic countries 220 | Uche Amazigo, Mounkaila Noma, Jesse Bump, Bruce Benton, Bernhard Liese, Laurent Yaméogo, Honorat Zouré, and Azodoga Seketeli in central Africa and East Africa. The program, spanning Participating countries, donors, and the partnership's 30 countries across Sub-Saharan Africa, encompasses more governing board have all endorsed the integration of than 107 projects to create a comprehensive approach to community-directed treatment into existing health systems eliminating the disease as a public health problem. The through other health interventions. Some countries, such as onchocerciasis control programs have yielded the following Uganda, have begun reorganizing their rural health services results: to use the community-based network as a national strategy. Some communities within the 30 African countries, along · 1989­90--60,000 people treated in 11 countries with NGDOs, such as Helen Keller International and The · 1994--2 million people treated Carter Center, have begun distributing bednets to prevent · 2002 (end of phase 1)--40 million people protected in malaria, and medications, including vitamin A, to prevent 11 countries; 600,000 cases of blindness prevented; malnutrition, pediatric blindness, and death; Praziquantel 18 million children spared the risk of onchocerciasis to control schistosomiasis; and ivermectin and albendazole · 2003--33 million people treated in 69,641 communities to control the transmission of lymphatic filariasis. in the APOC countries; more than 162,000 community distributors and 18,000 health workers trained or retrained · 2010--a projected 102 million people and about 100,000 REFERENCES communities protected in 16 countries and a projected Amazigo, U. 1993. "Onchocerciasis and Women's Reproductive Health: 150 million people protected in 30 countries. Indigenous and Biomedical Concepts." Tropical Doctor 23 (4): 149­51. ------. 1994. "Detrimental Effects of Onchocerciasis on Marriage Age Prior to 2002, when the OCP closed, 25 million hectares and Breastfeeding." Tropical and Geographical Medicine 46: 322­25. of relatively fertile land in the river valleys were freed for Awadzi, K., K. Y. Dadzie, H. Shulz-Key, D. R. Haddock, H. M. Gilles, and M. A. Aziz. 1985. "The Chemotherapy of Onchocerciasis X. An resettlement and for agriculture; the socioeconomic impact Assessment of Four Single Dose Treatment Regimes of MK-933 of OCP is considered to be enormous. More than 40 million (Ivermectin) in Human Onchocerciasis." Annals of Tropical Medicine people in the 11 countries were considered free from infec- and Parasitology 79 (1): 63­78. Benton, B. 1998. "Economic Impact of Onchocerciasis Control through tion and eye lesions. Sixteen million children born after the African Programme for Onchocerciasis Control: An Overview." 1974, when OCP activities began, are free of onchocerciasis; Annals of Tropical Medicine and Parasitology 92: S33­39. more than 1.5 million people originally infected are no Benton, B., J. Bump, A. Seketeli, and B. Liese. 2002. "Partnership and longer so; and more than 200,000 cases of blindness have Promise: Evolution of the African River-Blindness Campaigns." Annals of Tropical Medicine and Parasitology 96 (Supplement 1­March): 5­14. been prevented. Brieger, W. R., A. K. Awedoba, C. I. Eneanya, M. Hagan, D. Okello, K. F. The second phase of the studies to determine the impact Ogbuagu, O. O. Osanga, and E. B. L. Ovuga. 1998a. "The Effect of of APOC are under way, but the program is considered to be Ivermectin on Onchocercal Skin Disease and Severe Itching: Results of a Multi-Centre Trial." Tropical Medicine and International Health 3: highly cost-effective. The cost per disability-adjusted life 951­61. year of APOC operations is US$6.50. A study is also being Brieger, William R., Frederick O. Oshiname, and Oladele O. Ososanya. conducted to demonstrate the feasibility and cost-benefit of 1998b. "Stigma Associated with Onchocercal Skin Disease among Those Affected Near the Ofiki and Oyan Rivers in Western Nigeria." onchocerciasis elimination with ivermectin. Social Science and Medicine 47 (7): 841­852. The community-directed distribution network can get Dadzie, K. Y., J. Remme, E. S. Alley, and G. de Sole. 1990. "Changes in primary health care to the poorest of the poor through sim- Ocular Onchocerciasis Four and Twelve Months after Community- Based Treatment with Ivermectin in a Holoendemic Onchocerciasis ple, once- or twice-a-year interventions by nonmedical staff. Focus." Transactions of the Royal Society of Tropical Medicine and In some areas, additional activities have been carefully tested Hygiene 84: 103­8. and planned; in other areas, communities themselves took Dadzie, K. Y., J. Remme, G. de Sole, B. Boatin, E. S. Alley, O. Ba, and E. M. spontaneous action. Effective interventions exist, but they Samba. 1991. "Onchocerciasis Control by Large-Scale Ivermectin Treatment. Lancet 337: 1358­59. do not reach the people who need them most--the poor. Duke, B. O. L., and K. Y. Dadzie. 1993. Procedural Manual for Ivermectin Prior to the launching of APOC, not enough was known Distribution Programmes. Geneva: World Health Organization. about what works to reach the poor, that is, what systems Kale, O. O. 1998. "Onchocerciasis: The Burden of Disease." Annals of can ensure that poor and very remote populations have Tropical Medicine and Parasitology 92 (Suppl. 1): S101­15. Kim, A., and B. Benton. 1995."Cost­Benefit Analysis of the Onchocerciasis access to the benefits of health services, to available drugs, Control Program (OCP)." Technical Paper 282, World Bank, regardless of their locations. Washington, DC. Onchocerciasis | 221 Kim, A., A. Tandon, A. Hailu, H. Birrie, N. Berhe, A. Aga, G. Mengistu, et al. "Large Scale Ivermectin Distribution and Its Epidemiological 1997. "Health and Labour Productivity: The Economic Impact of Consequences." Acta Leidensia 59: 177­91. Onchocercal Skin Disease (OSD)." Policy Research Working Paper Richards Jr., F. O., B. Boatin, M. Sauerbrey, and A. Seketeli. 2001. "Control 1836, World Bank, Washington, DC. of Onchocerciasis Today: Status and Challenges." Trends in Parasitology Le Berre, R., J. F. Walsh, B. Philippon, P. Poudiougo, J. E. Henderickx, 17 (12): 558­63. P. Guillet, A. Seketeli, D. Quillevere, J. Grunewald, and R. A. Cheke. Vlassoff, C., M. Weiss, E. B. Ovuga, C. Eneanya, P. T. Nwel, S. S. Babalola, 1990. "The WHO Onchocerciasis Control Programme: Retrospect and A. K. Awedoba, B. Theophilus, P. Cofie, and P. Shatabi. 2000. "Gender Prospects." Philosophical Transactions of the Royal Society of London. and the Stigma of Onchocercal Skin Disease in Africa." Social Science Series B Biological Sciences 328 (1251): 721­27. and Medicine 50: 1353­54. Noma, M., B. E. Nwoke, I. Nutall, P. A. Tambala, P. Enyong, A. Namsenmo, Whitworth, J. A. G., C. E. Gilbert, D. M. Mabey, G. H. Maude, D. Morgan, J. Remme, U. V. Amazigo, O. O. Kale, and A. Seketeli. 2002. "Rapid and D. W. Taylor. 1991. "Effects of Repeated Doses of Ivermectin on Epidemiological Mapping of Onchocerciasis (REMO): Its Application Ocular Onchocerciasis: Community-Based Trial in Sierra Leone." by the African Programme for Onchocerciasis Control (APOC)." Lancet 338: 1100­03. Annals of Tropical Medicine and Parasitology 96 (Suppl. 1): S29­39. Whitworth, J. A., D. Morgan, G. H. Maude, A. J. Luty, and D. W. Taylor. Okello, D. O., E. B. Ovuga, and J. W. Ogwal Okeng. 1995. "Dermatological 1992. "A Community Trial of Ivermectin for Onchocerciasis in Sierra Problems of Onchocerciasis in Nebbi District, Uganda." East African Leone: Clinical and Parasitological Responses to Four Doses Given at Medical Journal 72: 295­98. Six-Monthly Intervals." Transactions of the Royal Society of Tropical Oladepo, O., W. R. Brieger, S. Otusanya, O. O. Kale, S. Offiong, and Medicine and Hygiene 86 (3): 277­80. M. Titiloye. 1997. "Farm Land Size and Onchocerciasis, Status of Peasant Farmers In South­Western Nigeria." Tropical Medicine and WHO (World Health Organization). 1973. PAG Mission: Controle de International Health 2: 334­40. l'Onchocercose dans la Région du Bassin de la Volta: Rapport de la Mission d'Assistance Preparatoire aux Gouvernements de Côte d'Ivoire, Ovuga, E. B., D. O. Okelo, J. W. Ogwal Okeng, N. Orwortho, and F. O. Dahomey, Ghana, Haute-Volta, Mali, Niger, Togo. Geneva: WHO. Opoka. 1995. "Social and Psychological Aspects of Onchocercal Skin Disease in Nebbi District of Uganda." East African Medical Journal 72: ------. 1987. WHO Expert Committee on Onchocerciasis, Third Report. 449­53. Technical Report Series 752. Geneva: WHO. Remme, J. H. F. 1995. "The African Programme for Onchocerciasis ------. 1995a. "Onchocerciasis and Its Control. Report of a WHO Expert Control: Preparing to Launch." Parasitology Today 11: 403­06. Committee on Onchocerciasis Control." World Health Organization Technical Report Series 852: 1­104. Remme, J., R. H. Baker, G. de Sole, K. Y. Dadzie, J. F. Walsh, M. A. Adams, E. S. Alley, and H. S. Avissey. 1989. "A Community Trial of Ivermectin ------. 1995b. The Pan-African Study Group on Onchocercal Skin Disease. in the Onchocerciasis Focus of Asubende, Ghana. I. Effect on the Report of a Multi-country Study. Document TDR/AFR/RP/95.1. microfilarial reservoir and the transmission of Onchocerca volvulus." Geneva: WHO. Tropical Medicine and Parasitology 40 (3): 367­74. ------. 1996. Report of a Multi-country Study on Community Directed Remme, J., G. de Sole, K. Y. Dadzie, E. S. Alley, R. H. A. Baker, J. D. F. Treatment with Ivermectin. Document TDR/AFR/RP/96.1. Geneva: Habbema, A. P. Plaisier, G. J. van Oortmarssen, and E. M. Samba. 1990. WHO. 222 | Uche Amazigo, Mounkaila Noma, Jesse Bump, Bruce Benton, Bernhard Liese, Laurent Yaméogo, Honorat Zouré, and Azodoga Seketeli Chapter 16 Maternal Mortality Khama O. Rogo, John Oucho, and Philip Mwalali At the close of the last century, Sub-Saharan Africa still chapter continues with a description of the levels and trends had high maternal morbidity and mortality rates, with the in maternal mortality in the decade 1990­2000. The causes goals of safe motherhood eluding many governments. The and correlates of maternal mortality, as well as priority Programme of Action of the International Conference on interventions, are examined. The last section of the chapter Population and Development of 1994 and the Fourth World points to what Sub-Saharan African countries could do Conference on Women of 1995 were created in an attempt to meet the maternal health component of the Millennium to tackle these issues and drew unprecedented attention Development Goals. to reproductive health and rights as well as to gender equity and equality. The scourge of the human immunodeficiency virus and acquired immune deficiency syndrome (HIV/ MEASUREMENT AND DATA SOURCES AIDS) has ravaged the region's population and has left in its wake untold destruction in the demographic, economic, Measuring maternal mortality remains one of the more dif- and social spheres (UN 2003). Demographic events of the ficult issues in maternal health, and yet an accurate picture last decade are a sharp contrast to those in the 1980s, when of the scope of the problem is important to implement decreasing infant, child, and adult mortality rates and approaches to improve maternal health care. maternal mortality ratios (MMRs) were leading to steadily increasing life expectancy and improved health status for Definitions women in the region. Data sets assembled since the 1990s are the basis for the The World Health Organization's (WHO's) 10th revision of analysis of maternal mortality in Sub-Saharan Africa in this the International Statistical Classification of Diseases and chapter. Beginning with an examination of measurement Related Health Problems (ICD-10) defines maternal mortality approaches and data sources of maternal mortality, the as "the death of a woman while pregnant or within 42 days 223 of termination of pregnancy irrespective of the duration Second, the maternal mortality rate is the number of and the site of the pregnancy, from any cause related to maternal deaths in a given period per 100,000 women of or aggravated by the pregnancy or its management but reproductive age during the same period: not from accidental or incidental causes" (WHO 1992). Furthermore, the ICD-10 introduced a new category, Maternal mortality rate namely, late maternal death, that is defined as "the death of No. of maternal deaths 100,000 (16.2) a woman from direct or indirect obstetric causes more than No. of women age 15­49 42 days but less than one year after termination of preg- The maternal mortality rate is a cause-specific mortality rate nancy." The ICD-10 defined direct obstetric deaths as for women of reproductive age in the presence of other "maternal deaths resulting from obstetric complications of causes of death. the pregnant state (pregnancy, labor, and the puerperium), Third, the lifetime risk of maternal death is the risk a from interventions, omissions, incorrect treatment, or from woman has of dying during her reproductive years, given a chain of events resulting from any of the above." Indirect current rates of fertility and the risk of maternal mortality. obstetric deaths, by contrast, are "those resulting from previ- Given the length of the reproductive period (about 35 years), ous existing disease or disease that developed during the lifetime risk is calculated as [1 (1 maternal mor- pregnancy and which was not due to obstetric causes, but tality rate)35] (AbouZahr and Wardlaw 2003). was aggravated by physiologic effects of pregnancy." Because accidental deaths are excluded from the definition of Data Sources maternal deaths, the ICD-10 introduced the term pregnancy-related death, defined as "the death of a woman Estimation of maternal mortality indicators is difficult and while pregnant or within 42 days of termination of preg- subject to error because the data on which the estimates are nancy, irrespective of the cause of death" (WHO 1992). based are frequently inaccurate. In the estimates produced Maternal mortality is associated with neonatal mortality, jointly by the WHO, the United Nations Children's Fund but a lack of data, especially for Sub-Saharan Africa, con- (UNICEF), and the United Nations Population Fund strains conclusive findings. A review of the literature sug- (UNFPA) (AbouZahr and Wardlaw 2003), countries are gests that a relationship exists between maternal mortality classified into one of the following four categories: and perinatal mortality, including stillbirths and early neonatal deaths, and that interventions that save the lives of 1. countries with complete civil registration and good mothers are also effective in reducing neonatal mortality cause-of-death attribution (Bang, Bang, and Reddy 2005; Darmstadt et al. 2005; 2. those with complete or nearly complete civil registration Jacobson 1991; Kwast 1996; Lompo et al. 1993; Onuh and of the number of births and deaths but with poor cause- Aisien 2004). of-death attribution 3. those without a reliable system of civil registration, where maternal deaths, like other vital events, go unrecorded Measures of Maternal Mortality 4. those with estimates of maternal mortality based on Three measures of maternal mortality are commonly used. household surveys, using direct or indirect sisterhood First, the maternal mortality ratio is expressed as the number methods. of maternal deaths during a given time period per 100,000 live births during the same period: Most Sub-Saharan African countries fall into the last two categories. This implies that estimates of maternal mortality No. of maternal deaths MMR 100,000 (16.1) in Sub-Saharan Africa are derived from the methods men- No. of live births tioned in category 4 above. When estimating maternal The MMR represents a measure of the risk of death once mortality for these countries, the WHO, UNICEF, and the a woman has become pregnant. As a ratio, it is not a true UNFPA use an approach that includes adjusting country risk, as it involves two different populations, pregnant data to account for underreporting and misclassification women and live newborns. The ratio can be influenced by and a statistical model to generate estimates. Maternal the prevalence of stillbirths as well as the prevalence of deaths are generally identified by medical certification when induced abortions. vital registration exists; in household surveys, censuses, and 224 | Khama O. Rogo, John Oucho, and Philip Mwalali Reproductive Age Mortality Studies (RAMOS), a time of or word of mouth. Generally, a team of physicians studies death definition is used, making these "pregnancy-related the information collected by the study team to determine deaths" rather than "maternal deaths." Several sources of maternal mortality (Stanton, Abderrahim, and Hill 1997). data on which estimates of maternal mortality rely are Investigation of maternal mortality by the sisterhood described briefly below. method, an indirect approach, entails asking respondents about ever-married sisters: how many have died, and how Vital Registration. This source is generally used in devel- many died while they were pregnant or during childbirth or oped countries, where maternal mortality is estimated from six weeks following the end of the pregnancy (Stanton, deaths registered by cause of death. It permits calculation of Abderrahim, and Hill 2000). This method has been used period-specific maternal mortality ratios in which the frequently in population-based approaches, such as house- numerator is registered maternal deaths and the denomina- hold surveys. Originated by W. J. Graham and W. Brass tor is the registered live births (see equation 16.1). Where the (Graham 1991) in a field trial in The Gambia in 1987, it is a degree of underreporting of maternal deaths is almost simple and low-cost technique with considerable appeal in similar to that of live births, the resulting maternal mortality maternal mortality studies. The method provides a retro- ratio will give a reasonable population-based estimate. In spective rather than a current estimate, averaging experi- Sub-Saharan Africa, vital registration is lacking in all coun- ence over a lengthy time period, 35 years, with a midpoint tries, with the exception of Mauritius. In the rest of Sub- of about 12 years before the survey. The estimates have Saharan Africa, vital registration-based maternal mortality wide confidence intervals, making short-term monitoring estimates are inadequate because of the failure to register of maternal mortality trends difficult (AbouZahr and vital events, as well as the misclassification of causes of Wardlaw 2003). death, due to the absence of medical personnel and for The direct sibling-history method has been employed by social, religious, and emotional reasons (Graham 1991). In the Demographic and Health Surveys (DHSs), and permits view of these shortcomings, the vital registration system is the calculation of a maternal mortality ratio for a more ruled out as a reliable source of measuring maternal mor- recent period of time; results are typically calculated for a tality in the region. reference period of seven years before the survey, with a point estimate some three to four years before the survey Population-Based Data. In the absence of complete vital (AbouZahr and Wardlaw 2003). The data requirements for registration, the main sources of population-based data are this method are more demanding than for the indirect censuses and surveys, which use one of the following approach because respondents are asked three sets of ques- approaches to estimate maternal mortality: RAMOS, sister- tions. First, they are asked how many children the mother hood method, sibling-history method, and household has given birth to and how many of the children were born deaths methods. The key characteristics of each of them in before the respondent. The second set of questions asks the context of Sub-Saharan Africa follow. about the siblings' sex, age, and survival, or how many years Decennial census counts can generate both national and ago she or he died and age at death. Finally, three questions subnational data with questions on deaths in the household are asked about dead sisters to determine maternal morta- in a defined reference period (one or two years before the lity: whether she was pregnant when she died, whether she census), followed by detailed questions that permit identifi- died during childbirth, and whether she died within two cation of maternal deaths according to time of death relative months after the end of pregnancy or childbirth (Stanton, to pregnancy. Estimates of maternal mortality derived from Abderrahim, and Hill 2000). censuses in this way are believed to be fairly accurate The final method entails undertaking a survey of house- (Stanton et al. 2001). A problem is that censuses of Sub- holds to ascertain maternal deaths. Unlike the sibling-history Saharan African countries ask questions that are often not method, in which a sibling reports on the deaths of female useful for identification of maternal deaths, which, in addi- siblings, the direct household method involves reporting by tion, are often not fully reported. household heads or any other persons volunteering such A RAMOS study seeks to identify all deaths of females of information. When household information is screened, reproductive age in a defined population by investigating all maternal and nonmaternal deaths can be distinguished. relevant sources of information, namely, household inter- Although useful in direct estimation of maternal mortality, views, hospital and health center records, vital registration, this type of household survey is expensive and often too Maternal Mortality | 225 complex to implement in many countries because it requires LEVELS AND TRENDS IN MATERNAL MORTALITY large sample sizes. An alternative approach is provided by demographic surveillance systems (DSSs) that exist in Estimates of maternal mortality based on survey data and selected districts in a few Sub-Saharan African countries, models have been compiled at the WHO for 1990, 1995, notably the Navrongo DSS in Ghana and the Nouna DSS in and 2000. Table 16.1 presents estimates of MMR and the Burkina Faso (http://www.indepth-network.org/dss_site_ total deaths relating to maternal mortality in the years 1990, profiles/dss_sites.htm). DSSs continually collect births and 1995, and 2000, but trends should be interpreted with deaths in households in the populations in the sites. caution, as the estimates are based on different models. Unfortunately, the DSSs generally cover small areas, Among regions, Sub-Saharan Africa had the highest MMR typically a district in a country, so that the maternal mortal- over the period 1990­2000. The wide margins of error ity estimates based on them do not permit generalization; (not shown in table 16.1) of the MMR estimates, and the they can be treated as estimates based on case studies. methodological approaches employed in the three years, limit analysis of trends over the period (AbouZahr and Health Services Data. Hospitals generally collect data on Wardlaw 2003). causes of deaths, including maternal deaths. But lower-level Table 16.2 shows maternal mortality measures (number facilities, such as clinics, dispensaries, or health posts, are of maternal deaths, MMR, and lifetime risk of maternal sometimes omitted in inquiries about maternal mortality. death) in Sub-Saharan Africa over the period 1990­2000. As a result, health services data in Sub-Saharan Africa are MMRs of 1,000 or more per 100,000 live births were record- often incomplete and misclassified, especially because ed in 16 of the Sub-Saharan Africa countries in 1990; that deaths related to ectopic pregnancy and abortion are level of MMR was recorded in 21 countries in 1995, and recorded in female wards rather than maternity wards. In 17 countries in 2000 (table 16.2). Many of the countries addition, deaths outside health facilities are generally with civil war or unstable governments--including Angola, excluded from the health services data. In Sub-Saharan Burundi, Central African Republic (since 1995), Chad, Africa, where 58 percent of deliveries take place outside of Eritrea, Ethiopia, Mozambique (except for 1995), Niger, health facilities, maternal mortality using health services Nigeria, Rwanda, Sierra Leone, and Somalia--fell into that data is underestimated (WHO 2005). category in the 1990s. Kenya and Tanzania, although not Table 16.1 Maternal Mortality Measures, 1990, 1995, and 2000 MMR (per 100,000 live births) Maternal deaths (thousands) Region 1990 1995 2000 1990 1995 2000 World total 430 400 400 585 515 529 Developed regions 27 21 20 4 2.8 2.5 Europe 36 28 28 3.2 2.2 1.7 Developing regions 480 440 440 582 512 527 Africa 870 1,000 830 235 273 251 North Africa 340 200 130 16 7.2 4.6 Sub-Saharan África 823a 1,100 920 219a 265 247 Asia 390 280 330 323 217 253 East Asia 95 55 55 24 13 11 South-central Asia 560 410 520 227 158 207 Southeast Asia 440 300 210 56 35 25 Western Asia 320 230 190 16 11 9.8 Latin America and the Caribbean 190 190 190 23 22 22 Oceania 680 260 240 1.4 0.6 0.5 Sources: AbouZahr and Wardlaw 2001, 2003; WHO and UNICEF 1996. Note: The estimates are based on models, which were specified differently for each of the years. The data are therefore not strictly comparable, and trends should be interpreted with caution. a. Average for Sub-Saharan Africa subregions: eastern, middle, southern, and western. 226 | Khama O. Rogo, John Oucho, and Philip Mwalali Table 16.2 Maternal Mortality Measures in Sub-Saharan Africa, by Country, 1990­2000 Lifetime risk of MMR maternal death No. of maternal deaths (per 1,000 births) (1 in the numbers below) Country 1990 1995 2000 1990 1995 2000 1990 1995 2000 1. Angola 7,200 7,100 11,000 1,500 1,300 1,700 8 9 7 2. Benin 2,300 2,000 2,200 990 880 850 12 15 17 3. Botswana 120 250 50 250 480 100 65 38 200 4. Burkina Faso 4,000 6,700 5,400 930 1,400 1,000 14 7 12 5. Burundi 3,400 5,100 2,800 1,300 1,900 1,000 9 21 12 6. Cameroon 2,600 3,800 4,000 550 720 730 26 21 23 7. Cape Verde .. 20 20 .. 190 150 .. 120 160 8. Central African Republic 850 1,500 1,600 700 1,200 1,100 21 14 15 9. Chad 3,700 4,500 4,200 1,500 1,500 1,100 9 9 11 10. Comoros 260 130 130 950 570 480 12 29 33 11. Congo, Democratic Republic of 16,000 20,000 24,000 870 940 990 14 13 13 12. Congo, Republic of 890 1,300 690 890 1,100 510 15 12 26 13. Côte d'Ivoire 4,900 6,000 3,900 810 1,200 690 14 13 25 14. Djibouti 110 120 180 570 520 730 24 29 19 15. Equatorial Guinea 130 240 180 820 1,400 880 17 10 16 16. Eritrea 1,900 1,600 930 1,400 1,100 630 10 12 24 17. Ethiopia 33,000 46,000 24,000 1,400 1,800 850 9 7 14 18. Gabon 210 250 200 500 620 420 32 25 37 19. Gambia, The 460 500 270 1,100 1,100 540 13 14 31 20. Ghana 4,800 4,000 3,500 740 590 540 18 26 35 21. Guinea 4,700 3,600 2,700 1,600 1,200 740 7 12 18 22. Guinea-Bissau 380 420 590 910 910 1,100 16 15 13 23. Kenya 7,000 13,000 11,000 650 1,300 1,000 20 13 19 24. Lesotho 420 370 380 610 530 550 26 32 32 25. Liberia 690 1,100 1,200 560 1,000 760 22 12 16 26. Madagascar 2,800 3,400 3,800 490 580 550 27 25 26 27. Malawi 2,700 2,800 9,300 560 580 1,800 20 21 7 28. Mali 5,700 3,000 6,800 1,200 630 1,200 10 19 10 29. Mauritania 750 850 1,200 930 870 1,000 16 17 14 30. Mauritius 25 10 5 120 45 24 300 880 1,700 31. Mozambique 9,800 7,400 7,900 1,500 980 1,000 9 13 14 32. Namibia 190 210 190 370 370 300 42 44 54 33. Niger 5,100 4,300 9,700 1,200 920 1,600 9 13 7 34. Nigeria 44,000 45,000 37,000 1,000 1,100 800 13 14 18 35. Reunion .. 5 5 .. 39 41 .. 930 970 36. Rwanda 400 6,300 4,200 1,300 2,300 1,400 9 6 10 37. Senegal 3,900 4,100 2,500 1,200 1,200 690 11 12 22 38. Sierra Leone 3,600 4,200 4,500 1,800 2,100 2,000 7 6 6 39. Somalia 7,000 7,100 5,100 1,600 1,600 1,100 7 7 10 40. South Africa 2,700 3,600 2,600 230 340 230 85 70 120 41. Swaziland 160 130 120 560 370 320 29 45 49 42. Tanzania 8,700 13,000 21,000 770 1,100 1,500 18 14 10 43. Togo 1,000 1,700 1,000 640 980 570 20 13 26 44. Uganda 11,000 10,000 10,000 1,200 1,100 880 10 11 13 45. Zambia 3,500 3,100 3,300 940 870 750 14 17 19 46. Zimbabwe 2,300 2,200 5,000 570 610 1,100 28 33 16 Sources: AbouZahr, Wardlaw, and Hill 2001; WHO and UNICEF 1996. Note: .. negligible. Countries for which data are not available for two of the three years are excluded. The estimates are based on models, which were specified differently for each of the years; the data are therefore not strictly comparable, and trends should be interpreted with caution. Maternal Mortality | 227 having suffered a conflict, are notable for maintaining occurred during pregnancy, 48 percent during childbirth, maternal mortality ratios above 1,000 maternal deaths per and 36 percent postpartum (Ghebrehiwot 2004). These 100,000 live births since 1995. Mauritius and Reunion in findings imply that the causes of the deaths in this critical the Indian Ocean, and Cape Verde in the Atlantic Ocean, period are either the result of labor or worsened by labor and the landlocked country of Botswana have low maternal and delivery. mortality ratios, atypical of Sub-Saharan Africa. As noted earlier, the causes of maternal mortality have traditionally been classified as direct and indirect, although the distinction is not always easy to discern (both CAUSES AND CORRELATES OF MATERNAL were grouped under pathogenic causes in the previous MORTALITY edition of this book). Pathogenic causes are purely medical and therefore best determined by health professionals. Most Understanding the causes and correlates of maternal mor- of the information on pathogenic causes is derived from tality is crucial in confronting the challenge of unyielding hospital studies; thus, data from health institutions will high rates in Sub-Saharan Africa. Abraham Lilienfeld, a continue to be an important source of information for prominent epidemiologist, very appropriately remarked, direct and indirect causes of maternal deaths. Implicit is "the better we know about the root cause of a problem, the the need to educate health professionals on the ICD and better we are in a position to address the problem," and in his provide updates whenever the ICD definition changes. As an book, Foundations of Epidemiology, cites Benjamin Disraeli's example, the 10th revision of ICD has introduced a much statement, "The more extensive a man's knowledge of what broader definition of maternal death and has expanded on has been done, the greater will be his power of knowing the categorization of the causes (WHO 1992). This will what to do" (Lilienfeld 1980). Despite the 1978 Alma Ata make analysis of trends increasingly more difficult because Declaration and the 1987 Safe Motherhood Initiative, the past data will need to be adjusted to accommodate the new reduction of maternal mortality has been minimal world- definition in order to make them comparable with more wide. The slow improvement in the MMR in the developing recent data. world is due not only to the trend in Sub-Saharan Africa but Availability and accuracy of data sources influence the also to stagnating declines in the regions of Latin America study of causes and correlates. For instance, data from hos- and the Caribbean and South Central Asia. In other regions, pitals or health institutions are limited in that medically notably North Africa and Southeast Asia, the MMR is esti- certified deaths at these institutions involve only a small and mated to have declined substantially during the 1990s. Such selective fraction of total deaths. This limitation is greatest divergent trends call for a closer examination of the factors in Sub-Saharan Africa, where a large proportion of deliver- correlated with the MMR. ies take place at home (WHO 2005). Different interactive factors contribute to maternal The main direct causes of maternal deaths, accounting for morbidity and mortality. The range is wide and includes the up to 80 percent of cases in Africa, are obstetric hemorrhage, behavior of families and communities, social status, educa- puerperal sepsis, pregnancy-induced hypertension (includ- tion, income, nutritional status, age, parity, and availability ing eclampsia), obstructed labor and ruptured uterus, and of health services. It is important to note that non­health complications of unsafe abortion (see figure 16.1). Three sector activities, such as education, water and sanitation, causes--hemorrhage, sepsis, and eclampsia--account for a roads and communication, agriculture, and internal vast majority of deaths, considering that even some cases of security, also influence maternal outcome. In Sub-Saharan abortion or obstructed labor eventually succumb to either Africa, some of the highest MMRs have been recorded in bleeding or sepsis. countries that are in conflict or have large refugee popula- Indirect causes account for 20 to 25 percent of maternal tions, such as Angola and Sierra Leone. deaths and are attributable to illnesses aggravated by preg- nancy (WHO 2005). They include anemia; malaria; HIV/AIDS; diseases of the heart, lung, liver, or kidneys; and Causes of Maternal Deaths ectopic pregnancies. Physical violence and accidents are not About 60 percent of the maternal deaths occur during child- included in this group. birth and the immediate postpartum period, with 50 percent As documented by several DHS surveys, many African of these deaths occurring within the first 24 hours of delivery. women enter pregnancy in a state of nutritional deficit and In a recent study in Eritrea, 16 percent of maternal deaths therefore are unprepared to cope with the extra physiological 228 | Khama O. Rogo, John Oucho, and Philip Mwalali Figure 16.1 Major Causes of Maternal Mortality in 46 million induced abortions globally every year, about Sub-Saharan Africa 20 million are considered unsafe. It is estimated that 95 per- Other causes cent of unsafe abortions occur in the developing world 8% (Henshaw, Singh, and Haas 1999). The WHO (1998) esti- Hemorrhage mated that there were about 5 million induced abortions in 25% Africa annually, whereas Rogo (1993), using the results of several DHS surveys, estimated that there were 1.5 million induced abortions, most of which were unsafe. The tragedy Abortion 33% of abortion-related mortality in Africa is that most of the victims are teenagers. The unsafe abortion conundrum in Africa begins with Sepsis 15% unprotected sex among teenagers who are ill-informed about their sexuality; an unwanted or ill-timed pregnancy Eclampsia follows. Living in countries where induced abortions are Obstructed labor 6% 13% legally restricted, the young victims resort to back street abortionists or quacks. Crude methods used in the pregnancy Source: WHO 1992. termination, delay in seeking medical attention when and if there is a problem, and the poor quality of postabortion care demands of pregnancy. In Eritrea, for example, 37.3 percent lead to a significant proportion of the victims sustaining of women have a low body mass index, which is an indica- serious injuries with life-threatening complications, result- tor of chronic energy deficiency (Eritrea National Statistics ing in either death or disability. For survivors the psycho- and Evaluation Office and ORC Macro 2002). The nutri- logical impact is immense and lifelong (Rogo 2004). tional deficit, macro- or micronutrient, predisposes these Postabortal sepsis is worse in HIV/AIDS-infected women women to anemia in pregnancy, among other problems. (Mbaruku 2005). Anemia is highly prevalent in Africa, with up to three-fifths of pregnant women in Africa having some degree of anemia, Determinants of Maternal Mortality and Morbidity and about one-third classified as having severe anemia (Isah et al. 1985; Massawe et al. 1996; Massawe et al. 1999; Van den Available evidence indicates that there are several factors Broek and Letsky 2000). Anemia may cause death on its own that predispose a woman to greater risk of maternal death. or predispose a woman to severe postpartum hemorrhage The common biomedical approach to the determinants of leading to death (Harrison 1997). maternal morbidity and mortality usually divides them into The growing HIV/AIDS pandemic is also having a severe distal and proximal factors. impact on women's health. It is estimated that there were The seminal work by McCarthy and Maine (1992) is 5 million new HIV infections in 2003, of which 40 percent credited with the conceptual model of analyzing determi- were among women and 20 percent among children (United nants of maternal mortality that could be applied to Nations 2003). In eastern and southern Africa, between 20 research as well as programs. The concept grouped the and 30 percent of pregnant women are infected with HIV, determinants as: and available evidence indicates that HIV/AIDS currently accounts for at least 18 percent of maternal deaths. Death in · distant, or socioeconomic, factors; this case results from opportunistic infections, puerperal · intermediate factors (health behavior and status, access sepsis, meningitis, tuberculosis, pneumonia, postabortion to services, and unknown factors); sepsis, encephalitis, and probably malaria (Mbaruku 2005; · outcomes (pregnancy, morbidity, and mortality). Pattinson et al. 2005). Unsafe abortion deserves special mention in Africa, The McCarthy and Maine concept has since been modi- the only region where complications of abortion are the fied, most notably by UNICEF (1999), to facilitate strategic most common cause of maternal mortality. Globally, unsafe programming for maternal health. From the pediatric abortion accounts for about 13 percent of maternal deaths perspective, the Mosley and Chen (1984) framework for the compared with 30 to 50 percent in Sub-Saharan Africa (AGI study of child survival in developing countries has also 1999; Henshaw, Singh, and Haas 1999). Of the estimated found, with various modifications, utility in the analysis of Maternal Mortality | 229 determinants of maternal morbidity and mortality. The group to another and cover a wide range of activities and original model proposed three levels of determinants of practices; from the sexual or genitally linked ones, such as child mortality (socioeconomic determinants, proximate female genital cutting, to feeding and nutritional practices. and biological determinants, and outcomes expressed in In addition, a plethora of harmful beliefs and practices terms of growth and death), but subsequent modifications around pregnancy and childbirth affect health-seeking have expanded the levels to five: household characteristics behavior during pregnancy and parturition. The dispropor- (behavioral), intermediate variables (behavioral and bio- tionately low use of health facilities for delivery care is testi- logical), risk factors (biological), malnutrition-infection mony to the strength of these beliefs (Ghebrehiwot 2004). syndrome, and demographic outcome (van Norren and van Household poverty, allocation of resources, and the con- Viannen 1986). The Poverty Reduction Strategy approach trol of those resources also influence maternal mortality. developed by the World Bank and sector-wide approaches Delivery of infants is not free of charge in many African to the health sector have generated new interest for incorpo- countries. Indeed, it was never without cost in traditional rating government policies and actions, within or outside societies either. Even in countries where delivery is declared the health sector, that focus on health outcomes. Edwards to be free in public facilities, the cost of accessing care, both (2001), by expanding on previous models, introduced the direct and indirect, can be prohibitive, quality notwith- macroeconomic evaluation of non­health sector policies standing. The relationship to poverty is bi-directional; that influence health. complications of pregnancy were cited as one of the most These developments are relevant to maternal health and common causes of household poverty (Borghi et al. 2003; can be applied to generating a more comprehensive under- Claeson et al. 2001). standing of determinants and correlates of maternal health in Africa. The following modified framework is proposed as Biological-Demographic Variables and Risk Factors. appropriate for discussing the correlates of maternal Standard biological variables, such as age, height, and parity, mortality in Africa: apply to maternal mortality in Africa as elsewhere. In many · household and community characteristics (behavior, countries of Sub-Saharan Africa, at least 50 percent or cultural-religious values, and income poverty) more of women will have started childbearing by age 19. · biological-demographic variables and risk factors Adolescents comprise about 20 percent of maternal deaths, · malnutrition-infection syndrome (including protein- most of which are due to complications of unsafe abortion. energy malnutrition [PEM], micronutrient deficiencies, Early marriage and childbearing are associated with high anemia, malaria, and HIV/AIDS) parity and therefore higher risk of maternal death · health systems (Ghebrehiwot 2004). Various indicators of maternal status · national policies and related investments (health and during pregnancy and childbirth may also be predictors of nonhealth). maternal outcome, including edema, hypertension, and history of previous complications (Garenne et al. 1997). Sociodemographic factors are correlates of maternal mor- Household and Community Characteristics. Pregnancy tality. Marital status, first pregnancy, and level of education outcome and maternal survival have strong correlations are commonly cited (Garenne et al. 1997). with household behavior and decision making. Enlightened communities value their mothers and seek prompt attention at the earliest indication of problems. Low status of women Malnutrition-Infection Syndrome. Malaria remains a in the household and society as a whole, as exemplified by major killer of women in pregnancy and a leading indirect inequality in education, employment, property ownership, cause of maternal mortality. There are effective interven- participation, and decision making, is another important tions, such as intermittent preventive treatment and correlate (Wall 1998). Gender-based violence is common in insecticide-treated bednets that are affordable but often not situations in which the status of women is low and legal available where they are most needed. The changing com- protection inadequate, and in turn it is correlated with high plexities of malaria chemotherapy and the rising cost of rates of maternal mortality. newer, more effective combinations pose new challenges, Harmful traditional practices and religious beliefs also including safety in pregnancy (Heymann et al. 1990; adversely affect maternal health. They vary from one ethnic Shulman et al. 1999). 230 | Khama O. Rogo, John Oucho, and Philip Mwalali HIV/AIDS and its effect on maternal outcomes in Africa Figure 16.2 Relation between Skilled Attendant at Delivery is grossly underreported. HIV is not regarded as a primary and MMR for All Countries, 1995 cause of death unless AIDS is diagnosed. A study in South 2,500 Africa reported a 25 percent increase in seropositivity, from 50 to 75 percent between 1997­99 and 2000, in maternal 2,000 deaths due to non­pregnancy-related sepsis in Pretoria 1,500 (Pattinson and Moodley 2002). HIV infection in pregnancy is also associated with anemia and severe malaria infections MMR 1,000 (Antelman et al. 2000). R 2 0.6124 As previously mentioned, both PEM and micronutrient 500 deficiencies are prevalent in African women. Pregnancy aggravates the situation and increases vulnerability to any 0 20 40 60 80 100 concurrent condition or opportunistic infection. Paul % skilled attendants at delivery (1993), in analyzing maternal mortality in Africa from 1980 Country, n 170 to 1987 found a strong correlation with calorie supply as a Source: Safe Motherhood Initiative and Maternal Mortality in 1995. Estimates developed percentage of requirements. Maternal anemia, however by WHO, UNICEF, and UNFPA 2001. mild, also increases several-fold the risk of life-threatening postpartum hemorrhage. implying no change from the 1990 global average (WHO 2005). Health Systems. Poorly financed and unaccountable Although there is no specific comparative rate, in its health systems, including weak referral systems, are a key global estimates on births attended by skilled personnel the determinant of maternal outcome. Another determinant is WHO (2005) reported 46.2 percent for Africa with the poor access to quality maternal health care services because lowest rates in East Africa (32.5 percent) and West Africa of geographical terrain and poor roads. Maternal health care (39.7 percent) (http://www.who.int/reproductive-health/ services are deemed to be of poor quality if, for example, global_monitoring/skilled_attendant.html). they lack skilled health providers, the providers have nega- Figure 16.2 shows that the higher the proportion of deliv- tive attitudes, treatment guidelines and protocols are inap- eries with a skilled attendant in a country, the lower the propriate, and they lack essential drugs, equipment, and country's MMR. Furthermore, most of the Sub-Saharan supplies. A low health personnel-to-population ratio is a Africa countries (not labeled) are above the regression line. chronic issue in Sub-Saharan Africa. For instance, the health Lack of or poorly functioning health management informa- personnel-to-population ratio in Sub-Saharan Africa is tion systems with an effective feedback loop as well as weak reported as 1:23,540, ranging from 1:750 in South Africa to supervision are further challenges influencing the quality of 1:72,000 in Rwanda. For nurses, the Sub-Saharan African maternal services and MMRs. health personnel-to-population ratio is 1:3,460, ranging from 1:600 in Zambia to 1:5,470 in Tanzania (Howson, National Policies and Investments. For any program or Harrison, and Law 1996). strategy on maternal health and safe motherhood to Given that skilled birth attendants working within a succeed, it must have the support of the highest level of supportive health system are the most important factor in national authority. Such support facilitates the allocation of keeping women healthy and safe in pregnancy, inadequate adequate financial and human resources; improves the numbers and distribution of human resources are a major infrastructure and communications; and puts in place effec- underlying cause of maternal mortality in Sub-Saharan tive and implementable standards, policies, and protocols. Africa. Although the use of skilled attendants at delivery Most countries in Sub-Saharan Africa have not addressed increased significantly in the developing world as a whole, policy issues, even where the policies have been shown to from 41 percent in 1990 to 57 percent in 2003, the greatest have significant influence on maternal mortality. Romania improvements were in Southeast Asia and North Africa provides the best example for the developing world of the and the least in Sub-Saharan Africa (http://unstats.un.org/ impact of changes in policy. Figure 16.3 clearly demon- unsd/mi/goals_2005/goal_5.pdf). A recent WHO report strates the trends of maternal mortality that occurred with indicates an average of 42 percent in the Africa Region, the change in the country's abortion law in 1966 and 1989 Maternal Mortality | 231 Figure 16.3 Effects of the Introduction in Romania of an Anti- REDUCING MATERNAL MORTALITY: PRIORITY Abortion Law in 1966 and Legalization of Abortion in 1989 INTERVENTIONS AND LESSONS 180 Many lessons have been learned on what works in maternal 160 Maternal deaths per 100,000 live births health. These have led to the identification of key interven- 140 tions for the reduction of maternal morbidity and mortality births 120 in the developing world. live 100 1,000 80 per 60 Priority Interventions Abortion deaths per 100,000 live births deaths 40 The past two decades have witnessed significant shifts in 20 thinking about effective interventions for improving mater- 0 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 nal outcomes in poor countries, from the Maternal and Child Health program to the Safe Motherhood Initiative, year with its "Pillars," and the Making Pregnancy Safer program Source: WHO 1998. (Starrs 1998; http://w3.whosea.org/pregnancy/chap1f.htm). More recently, the case for identifying and investing in the (http://www.who.int/docstore/world-health-day/en/ most effective interventions for safe motherhood has domi- pages1998/whd98_10.html). In the 1950s and early 1960s, nated the debates (AbouZahr, Wardlaw, and Hill 2001). the law provided for access to abortion and was associated There is evidence of effective clinical interventions that save with relatively low mortality ratios. The restrictions on lives, but less is known about the best strategies for acceler- abortion that followed was associated with significant ating reduction of maternal mortality in developing coun- increase in the MMR in the 1980s. Immediately after the tries, especially in Africa and South Asia (De Brouwere and December 1989 revolution that overthrew President Nicolae Van Lerberghe 2002; Donnay 2002; Koblinski, Campbell, Ceaus¸escu, restrictions on contraceptives were removed and and Heichelheim 1999; Liljestrand 2000). The findings of abortion legalized. Subsequently, with the precipitous drop the World Bank's analysis for the decline in maternal mor- in abortion-related mortality, the MMR dropped. tality in Malaysia and Sri Lanka in the past 50 to 60 years Changing the abortion policy reduced maternal mor- and the magnitude of health system expenditures on mater- tality by more than half in less than 10 years. Therefore, by nal health are relevant to Africa (World Bank 2003). changing the underlying policy-related causes, Sub-Saharan Malaysia and Sri Lanka have succeeded in reducing African countries have the potential of achieving reduc- maternal mortality to levels comparable to those in indus- tions in maternal mortality. In most Sub-Saharan African trial countries in the last few decades. Expanded female lit- countries, despite many international pronouncements, eracy in Sri Lanka and strong economic performance by high-level support for maternal health and measures to Malaysia helped promote these gains. The World Bank reduce maternal mortality and unsafe abortion is weak or analysis confirmed that maternal mortality can be halved in nonexistent. developing countries every 7 to 10 years and is affordable, Inadequate financing and sustainability of the health sec- regardless of income level and economic growth rate; steady, tor in general and of reproductive health in particular, are modest investment in poverty reduction and in maternal other barriers. In most African countries, health expendi- health services to improve access to and quality of emer- tures have not increased substantially while major problems gency obstetric care are required. in allocation efficiency and inequities exist (World Bank Removal of financial barriers to maternal care was an 2005). With various competing priorities for a dwindling important step in both countries, as was increased access to financial resource base, the health sector needs to do a bet- skilled birth attendance and emergency obstetric care. ter job in reclaiming its rightful share. Moreover, given the Recording and reporting of maternal deaths was a prerequi- inadequate investment, the number of health personnel site to addressing the challenges of reducing maternal trained is often small, and once trained, many open private mortality in both countries. Other important lessons were clinics or emigrate to developed countries to earn a better that governments can afford to provide the critical elements living. of maternal care free of charge to clients and that different 232 | Khama O. Rogo, John Oucho, and Philip Mwalali tactics are needed at different stages of the development of health systems. The transition from high to low MMR passes Box 16.1 Model of Three Levels of Delay through several phases characterized by the following: Level 1 delay: decision making at community · high MMR: low levels of skilled attendance and emer- level--examines decision-making process on preg- gency obstetric care (EmOC) nancy and childbirth at household and community · declining MMR: medium levels of skilled attendance and level, including birth preparedness EmOC Level 2 delay: accessibility, transport, and commu- · low MMR: high levels of skilled attendance and EmOC. nication--examines options for communication and transport from community to a health facility Except for South Africa and Botswana, most of Sub- Level 3 delay: availability of appropriate care, Saharan Africa falls in the first category, "high MMR--low including quality--examines perceived and actual levels of skilled attendance and EmOC." This status calls for quality of care provided to the client on arrival at establishing a solid foundation for effective maternity care, the facility increasing access to care, and ensuring appropriate use of Source: Thaddeus and Maine 1994. available services through community mobilization and improved quality. Elements of the foundation to support effective maternal care in Malaysia and Sri Lanka included professionalization of midwifery, civil registration of births, Delay" approach to analyze maternal mortality in Eritrea, compilation of data on maternal deaths, and replication of Ghebrehiwot (2004) attributed the causes of death to the local success. These elements do not always need additional following processes: resources but require focused leadership and effective management. · Delay One: failure or delay in recognition of danger signs--33 percent of maternal deaths · Delay Two: delay in deciding to seek care--40 percent of Improving Emergency Obstetric Care: The Three the cases Levels of Delay · Delay Three: delay in reaching appropriate care--19 per- It is clear from the foregoing that accelerating the decline of cent of cases maternal mortality in Sub-Saharan Africa and realization of · Delay Four: delay in receiving appropriate care--52 per- the Millennium Development Goals will require the provi- cent of cases. sion of a synergistic package of health and social services that reaches everyone, especially deprived populations. This simple framework appears ideal for Sub-Saharan The framework model of three delays has been applied to Africa. It works well and supports local partners in finding analyze the constraints, opportunities, and systems required tailor-made solutions to challenges posed in each specific at different levels of a safe motherhood program (box 16.1). setting and making service more responsive to local com- This framework serves as a useful planning tool for the munity needs. It can also be used to improve data collection actions required at every level of health care while empha- and use at the local level. sizing the need to link these levels through transport and The first level involves a primary health care bottoms-up communication, supervision, and community outreaches. approach with active community involvement (men and Thus, community awareness and trust and better access to women) and focused comprehensive development pro- and quality of emergency transport reinforce each other to grams wherein reproductive health and safe motherhood improve maternal outcome. In an integrated essential health are appropriately integrated into the district health system. care package, this network enhances provision of other The second level entails expanding access to quality services, such as family planning and immunization, while services, including functional linkages between communi- promoting emphasis on skilled attendance. ties and health facilities in regard to transport and commu- The model is advocated by the Regional Prevention of nication. This leads to the final level, where appropriate Maternal Mortality Network (PMMN 1995) and is rapidly quality of services is provided to clients on arrival at the gaining ground in Africa. Using a modified "Four Levels of health facility. Maternal Mortality | 233 A Tanzania case study (Urassa et al. 1997) depicts a typical the estimates are double those reported from country-based finding from most maternal mortality reviews in Africa. A studies, and underreporting was presented as the main large proportion of women die because of delayed decision factor. For instance, WHO, UNFPA, and UNICEF estimates making at home, lack of transport, and inappropriate care if reflect Kenya's MMR of 1,300 maternal deaths per 100,000 they make it alive to a health facility. This confirms the live births against the 590 maternal deaths per 100,000 live observation that reduction of high maternal mortality births reported by the DHS (AbouZahr and Wardlaw 2003). demands a strong focus on each level of delay through Also, the WHO, UNFPA, and UNICEF estimate for Tanzania creation of an effective system providing EmOC. Links was 1,500 maternal deaths per 100,000 live births, compared between the different levels of the health care system, from with the DHS estimate of 529 maternal deaths per 100,000 community through the basic health center (basic EmOC) live births (AbouZahr and Wardlaw 2003). These discrepan- to the referral hospital (comprehensive EmOC) are critical. cies are enormous and have significant implications in regard to the Millennium Development Goals. Increasing interest in the "near-miss population" pro- CONCLUSION vides an opportunity to explore the underlying causes and correlates of maternal morbidity and mortality. Near-miss Countries in Sub-Saharan Africa face the challenge of iden- population is defined as individuals who present to the tifying the base MMR to be reduced by the prescribed health facilities with life-threatening conditions or develop proportion in view of the conflicting MMR statistics from life-threatening complications while under management country-based studies and global estimates. Because of the (Filippi et al. 2000; Kaye et al. 2003; Mantel et al. 1998). This difficulties in assessing MMR, different methodologies have is a special category of survivors, whose stories provide been used. The use of different approaches complicates unique insights and valuable information on maternal the comparison and study of trends and causes of maternal mortality. In the absence of accurate and detailed informa- mortality, owing to variations in coverage, reference dates, tion on maternal deaths in Sub-Saharan Africa, increased and data presentation. use of studies on the near-miss population could provide Two Millennium Development Goals are of direct useful lessons. Such studies could provide information on relevance to maternal and newborn health: goal 4, to reduce the sequence of events leading to complications and child mortality, and goal 5, to improve maternal health. The describe critical life-saving interventions. Comparative latter goal is aimed at reducing the 1990 MMRs by 75 per- studies between the near misses and deaths occurring in the cent by 2015. The probability of reaching this goal in most same institution or communities could further clarify the Sub-Saharan Africa countries is highly questionable given factors contributing to or averting deaths. that most countries have shown little or no change in their MMRs and in a few countries they have even increased (World Bank 2004). Evidence from Sri Lanka and Malaysia REFERENCES shows that maternal mortality reduction can be accelerated through joint government and community action. For Sub- AbouZahr, C., and T. Wardlaw. 2001. "Maternal Mortality at End of a Decade: Signs of Progress." Bulletin of the World Health Organization Saharan Africa, this calls for the establishment of a solid 79: 32­34. foundation for effective maternity care and increasing access ------. 2003. "Maternal Mortality in 2000: Estimates Developed by and use of services. Monitoring of maternal care services WHO/UNICEF/UNFPA." WHO, Geneva. and audits of maternal deaths are crucial to this effort. AbouZahr, C., T. Wardlaw, and K. Hill. 2001. "Maternal Mortality in 1995: Estimates Developed by WHO, UNICEF, UNFPA." WHO/RHR/01.9, Further exploration into maternal mortality, especially WHO, Geneva. the underlying factors in Sub-Saharan Africa, is necessary, as AGI (Alan Guttmacher Institute). 1999. Sharing Responsibility; Women, are improved data collection systems, credibility in dissemi- Society and Abortion Worldwide. New York: AGI. nation of data, and realistic estimates where information is Antelman, G., G. I. Msamanga, D. Spiegelman, E. J. N. Urassa, R. Nahr, D. J. Hunter, and W. W. Fawzi. 2000. "Nutritional Factors and deficient. Without good sources of data, Sub-Saharan Africa Infectious Disease Contribute to Anemia among Pregnant Women will continue to rely on model estimates, which could be with HIV in Tanzania." Journal of Nutrition 130: 1950­57. misleading and de-motivating. For instance, the WHO, Bang, A. T., R. A. Bang, and H. M. Reddy. 2005. "Home-Based Neonatal UNICEF, and UNFPA estimates, the best and most quoted, Care: Summary and Applications of the Field Trial in Rural Gadchiroli, India 1993­2003." Journal of Perinatology 25: S108­122. vary significantly from those obtained from country-based Borghi, J., K. Hanson, C. Adjei Acquah, G. Ekanmian, V. Filippi, C. studies (AbouZahr and Wardlaw 2003). In some countries, Ronsmans, R. Brugha, E. Browne, and E. Alihonou. 2003. "Costs of 234 | Khama O. Rogo, John Oucho, and Philip Mwalali Near-Miss Obstetric Complications for Women and Their Families in Liljestrand, J. 2000. "Strategies to Reduce Maternal Mortality Worldwide." Benin and Ghana." Health Policy and Planning 18 (4): 383­92. Current Opinion in Obstetrics and Gynecology 12 (6): 513­17. Claeson, M., C. C. Griffin, T. A. Johnston, M. McLachlan, A. L. B. Soucat, Lompo, K., Y. J. Hutin, G. Traore, F. Tall, J. B. Guiard-Schmid, G. Yameogo, A. Wagstaff, and A. S. Yazbeck. 2001. Poverty Reduction and the Health and B. Fabre-teste. 1993. "Morbidity and Mortality Related to Sector. Washington, DC: World Bank. Obstetrical Referral Patients to the Hospital of Bobo-Dioulasso, Darmstadt, G. L., Z. A. Bhutta, S. Cousens, T. Adam, N. Walker, L. de Bernis, Burkina Faso." Annales de la Societe Belge de Medecine Tropicale 73 (2): and the Lancet Neonatal Survival Steering Team. 2005. "Evidence- 153­63. Based, Cost-Effective Interventions : How Many Newborn Babies Can Mantel, G. D., E. Buchmann, H. Rees, and R. C. Pattinson. 1998. "Severe We Save?" Lancet 365: 977­88. Acute Maternal Morbidity: A Pilot Study of a Definition for a Near- De Brouwere, V., and W. Van Lerberghe. 2002. "Safe Motherhood Miss." British Journal of Obstetrics and Gynaecology 105 (9): 985­90. Strategies: A Review of the Evidence." Studies in Health Services Massawe, S. N., E. Urassa, G. Lindmark, B. Moller, and L. Nystrom. 1996. Organization and Policy 17: 415­48. "Anemia in Pregnancy: A Major Health Problem with Implications for Donnay, F. 2002. "Maternal Survival in Developing Countries. What Can Maternal Health Care." African Journal of Health Sciences 3: 126­32. Be Done in the Next Decade?" International Journal of Gynecology and Massawe, S. N., E. N. Urassa, M. Mmari, G. Ronquist, G. Lindmark, and Obstetrics 70: 89­97. L. Nystrom. 1999. "The Complexity of Pregnancy Anemia in Dar es Edwards, R. T. 2001. "Paradigms and Research Programs: Is It Time to Salaam." Gynecologic and Obstetric Investigation 47: 76­82. Move from Health Care Economics to Health Economics?" Health Mbaruku, G. 2005. "Enhancing Survival of Mothers and Their Newborns Economics 19: 635­49. in Tanzania." PhD diss., Karolinska Institute, Stockholm. Eritrea National Statistics and Evaluation Office and ORC Macro. 2003. McCarthy, J., and D. Maine. 1992. "A Framework for Analyzing the Eritrea Demographic and Health Survey 2002. Calverton, MD: Eritrea Determinants of Maternal Mortality." Studies in Family Planning 23: National Statistics and Evaluation Office and ORC Macro. 23­33. Filippi, V., C. Ronsmans, T. Gandaho, W. Graham, E. Alihonow, and P. Mosley, W. H., and L. C. Chen. 1984. "An Analytical Framework for the Santos. 2000. "Women's Reports of Severe (Near-miss) Obstetric Study of Child Survival in Developing Countries." Population and Complications in Benin." Studies in Family Planning 31 (4): 309­24. Development Review 10: S24­45. Garenne, M., K. Mbaye, M. D. Bah, and P. Correa. 1997. "Risk Factors for Onuh, S. O., and A. O. Aisien. 2004. "Maternal and Fetal Outcome in Maternal Mortality: A Case-Control Study in Dakar Hospitals Eclamptic Patients in Benin City, Nigeria." Journal of Obstetrics and (Senegal)." African Journal of Reproductive Health 1 (1): 14­24. Gynecology 24 (7): 765­68. Ghebrehiwot, M. 2004. "Measurement of Maternal Mortality in Eritrea." Pattinson, R. C., and J. Moodley. 2002. Saving Mothers. Second Report on PhD diss., Johns Hopkins Bloomberg School of Public Health, Confidential Enquiries into Maternal Deaths in Pretoria, South Africa, Baltimore. 1999­2001. Department of Health, South Africa. Graham, W. J. 1991. "Maternal Mortality: Levels, Trends and Data Pattinson, R. C., H. Vandecruys, A. P. Macdonald, and G. D. Mantel. 2005. Deficiencies." In Disease and Health in Sub-Saharan Africa, ed. R. G. "Why Do Women Die during Childbirth?" Science in Africa, April 11. Feachem and D. R. Jamison, 101­16. New York: Oxford University Press. Paul, B. K. 1993. "Maternal Mortality in Africa: 1980­87." Social Science Harrison, K. 1997. "Maternal Mortality in Nigeria: The Real Issues." and Medicine 37 (6): 745­52. African Journal of Reproductive Health 1 (1): 7­13. PMMN (Prevention of Maternal Mortality Network). 1995. "Situation Henshaw, S. K., S. Singh, and T. Haas. 1999. "The Incidence of Abortion Analyses of Emergency Obstetric Care: Examples from Eleven Worldwide." International Family Planning Perspective 25: S30­38. Operations Research Projects in West Africa." Social Science and Medicine 40 (5): 657­67. Heymann, D. L., R. W. Steketee, J. J. Wirima, D. A. McFarland, C. O. Khoromana, and C. C. Campbell. 1990. "Antenatal Chloroquine Rogo, K. 1993. "Induced Abortion in Sub-Saharan Africa." East African Prophylaxis in Malawi: Chloroquine Resistance, Compliance, Medical Journal 70 (6): 386. Protective Efficacy and Cost." Transactions of the Royal Society of ------. 2004. "Improving Technologies to Reduce Abortion-Related Tropical Medicine and Hygiene 84: 496­98. Morbidity and Mortality." International Journal of Gynecology and Howson, C. P., P. F. Harrison, and M. Law. 1996. In Her Lifetime: Female Obstetrics 85: 552­61. Morbidity and Mortality in Sub-Saharan Africa. Washington, DC: Shulman, C. E., E. K. Dorman, F. Cutts, K. Kawuondo, J. Bulmer, and K. National Academy Press. Marsh. 1999. "Intermittent Sulphadoxine-Pyrimethamine to Prevent Isah, H. S., A. F. Fleming, I. A. O. Ujah, and C. G. Ekwempu. 1985. "Anemia Severe Anemia Secondary to Malaria in Pregnancy: Randomized and Iron Status of Pregnant and Non-Pregnant Women in the Guinea Placebo Controlled Trial." Lancet 353: 632­36. Savanna of Nigeria." Annals of Tropical Medicine and Parasitology Stanton, C., N. Abderrahim, and K. Hill. 1997. DHS Maternal Mortality 79 (5): 485­93. Indicators: An Assessment of Data Quality and Implications for Data Use. Jacobson, J. L. 1991. "Zimbabwe Birth Force." World Watch 4 (4): 5­6. DHS Analytical Reports 4. Calverton, MD: Macro International. Kaye, D., F. Mirembe, F. Aziga, and B. Namulema. 2003. "Maternal ------. 2000. "An Assessment of DHS Maternal Mortality Indicators." Mortality and Associated Near-Misses among Emergency Intrapartum Studies in Family Planning 31: 111­23. Obstetric Referrals in Mulago Hospital, Kampala, Uganda." East Stanton, C., J. Hobcraft, K. Hill, N. Kodjogbe, and W. T. Mapeta. 2001. African Medical Journal 80 (3): 144­49. "Every Death Counts: Measurement of Maternal Mortality Via Koblinski, M. A., O. Campbell, and J. Heichelheim. 1999. "Organizing Census." Bulletin of the World Health Organization 79 (7): 657­64. Delivery Care: What Works for Safe Motherhood?" Bulletin of the Starrs, A. 1998. The Safe Motherhood Action Agenda: Priorities for the Next World Health Organization 77 (5): 399­406. Decade. New York: Family Care International. Kwast, B. E. 1996. " Reduction of Maternal and Perinatal Mortality in Rural Thaddeus, S., and D. Maine. 1994. "Too Far to Walk: Maternal Mortality in and Peri-Urban Settings: What Works." Review. European Journal of Context." Social Science and Medicine 38 (8): 1091­110. Obstetrics, Gynecology, and Reproductive Biology 69 (1): 47­53. UNICEF (United Nations Children's Fund). 1999. Programming for Safe Lilienfeld, A. 1980. Foundations of Epidemiology. 2nd ed. New York: Oxford Motherhood: Guidelines for Maternal and Neonatal Survival. New York: University Press. UNICEF. Maternal Mortality | 235 United Nations. 2003. "Workshop in HIV/AIDS and Adult Mortality in ------. 1998. "Unsafe Abortion: Global and Regional Estimates of Developing Countries." UN/POP/2003/3, United Nations, New York. Incidence of and Mortality Due to Unsafe Abortion with a Listing of Urassa, E., S. Massawe, G. Lindmark, and L. Nystrom. 1997. "Operational Available Country Data." WHO/RHT/MSM/97.16, WHO, Geneva. Factors Affecting Maternal Mortality in Tanzania." Health Policy Plan ------. 2005. Reducing Maternal Deaths: The Challenge of the New 12 (1): 50­57. Millennium in the African Region. Brazzaville: WHO. Van den Broek, N. R., and E. A. Letsky. 2000. "Etiology of Anemia in WHO and UNICEF. 1996. Revised 1990 Estimates of Maternal Mortality: A Pregnancy in South Malawi." American Journal of Clinical Nutrition 72: New Approach by WHO and UNICEF. Geneva: WHO. 247S­56S. World Bank. 2003. Investing in Maternal Health: Learning from Malaysia Van Norren, B., and H. A. W. van Viannen. 1986. The Malnutrition and Sri Lanka. Washington, DC: World Bank. Infection Syndrome and Its Demographic Outcome in Developing ------. 2004. Making Services Work for Poor People. World Development Countries. Publication 4. The Hague: Programming Committee for Report. Washington, DC: World Bank. Demographic Research. ------. 2005. Improving Health, Nutrition, and Population Outcomes in Wall, L. L. 1998. "Dead Mothers and Injured Wives: The Social Context of Sub-Saharan Africa: The Role of the World Bank. Washington, DC: Maternal Morbidity and Mortality among the Hausa of Northern World Bank. Nigeria." Studies in Family Planning 29: 341­59. WHO (World Health Organization). 1992. International Statistical Classification of Diseases and Related Health Problems. 10th rev. ed. Geneva: World Health Organization. 236 | Khama O. Rogo, John Oucho, and Philip Mwalali Chapter17 HIV/AIDS Souleymane Mboup, Rosemary Musonda, Fred Mhalu, and Max Essex The acquired immune deficiency syndrome (AIDS) was DEVELOPMENT OF THE AIDS PANDEMIC first recognized as a disease in the early 1980s. Within about five years it became clear that a new epidemic of In retrospect, it is clear that several characteristics of unprecedented proportions was spreading throughout HIV/AIDS resulted in a serious underestimation of the Sub-Saharan Africa. Destruction of the immune system, importance of the epidemic by both individuals and the main characteristic of the disease, caused patients to societies. One feature of HIV/AIDS that is highly unusual die from a range of opportunistic infections. As the for an infectious disease is its consistently long incubation opportunistic infections that occurred reflected the period combined with a high rate of disease development. prevalence of given pathogens in the afflicted population, Most organisms cause clinical signs or symptoms in only a tuberculosis was one of the most common outcomes in fraction of those they infect. When infection occurs, as Africa. Soon after the recognition of AIDS, a new group in diseases such as measles or smallpox, the induction of retroviruses, subsequently designated the human period is short. HIV/AIDS is unique in that it regularly immunodeficiency virus (HIV), was identified as the causes lethal disease after a prolonged induction period that probable cause (Barre-Sinoussi et al. 1983; Gallo et al. lasts several years. As a result, the vast majority of HIV- 1984; Kitchen et al. 1984; Popovic et al. 1984). Some peo- infected people are clinically asymptomatic and do not ple were reluctant to accept the evidence that HIV was know they are infected unless they undergo serologic test- the cause of AIDS, both in the West and in Africa ing. Because HIV-infected asymptomatic people can trans- (Duesberg 1988). Many political leaders also chose to ig- mit the virus to others, the epidemic accelerates robustly. In nore or deny the importance of the expanding epidemic some settings, a significant fraction of the population until after widespread transmission of HIV had already becomes infected before disease and death have provided occurred. the most evident lesson of the need for vigilance. 237 HIV/AIDS is also difficult to control because it is a sexu- Figure 17.1 Disease Burden and Treatment of AIDS in ally transmitted disease. In most societies both leaders and Relation to Global Population and Economy citizens are reluctant to discuss sex. Although condoms are 90 efficient in preventing infection, they are often not an avail- 80 able option, for example, when couples wish to have chil- dren. The image of AIDS as a sexually transmitted disease 70 also contributes to the stigmatization of infected people, 60 which in turn may cause many to avoid testing to determine 50 their status. The variability in clinical outcomes also contributed to percent 40 the difficulty in controlling the spread of HIV, particularly 30 during the early stages of the epidemic. Tuberculosis, 20 Kaposi's sarcoma, and chronic diarrhea can all occur in the absence of HIV infection. In the absence of testing for HIV, 10 it is still possible for AIDS patients with such outcomes, and 0 Sub-Saharan Africa United States their families, to deny the infection and perhaps avoid some of the associated discrimination and stigma. As a societal % global population % global HIV/AIDS cases GNP as % global economy % AIDS cases on ARV problem, AIDS is also devastating because it usually attacks young adults during their most productive years of employ- Source: CIA 2006; UNAIDS 2004; UNAIDS and WHO 2005; WHO 2005. ment and parenting. Because both parents are often infected, AIDS epidemics ordinarily leave large numbers of Figure 17.2 Global Burden of HIV-1 Infection orphans behind. 25 20 RATES OF HIV/AIDS IN AFRICA 15 prevalence 10 In its most recent projections, the Joint United Nations % 5 Programme on HIV/AIDS (UNAIDS) has estimated that about 40 million people are currently infected with HIV, of 0 North America and Sub-Saharan Southern Africa All other developing whom about 25.8 million, or 64 percent of the total, are in Western Europe Africa countries outside Sub-Saharan Africa Sub-Saharan Africa (UNAIDS and WHO 2005). The esti- mates are that 4.9 million people became infected during Source: UNAIDS 2004. 2005, of whom 3.2 million, or 65 percent, were in Sub- Saharan Africa. During the same year, 77 percent of an esti- countries whose gross national productivity represents a mated global burden of 3.1 million AIDS deaths were pro- small fraction of the global economy (see figure 17.1). jected for Sub-Saharan Africa. During recent years, the use Wide variation in HIV prevalence rates also occurs with- of antiretroviral (ARV) drugs in the United States and other in Africa (Essex and Mboup 2002). However, overall adult developed countries has dramatically reduced AIDS death rates of HIV in Sub-Saharan Africa are about 7.2 percent, rates, but until now, only a small proportion of AIDS much higher than in other regions of the world, including patients in Sub-Saharan Africa have received treatment (see other regions with large populations living in developing figure 17.1). countries (see figure 17.2). Rates of HIV are low in North Several factors account for the low fraction of AIDS Africa, although countries that span the Sahara, such as patients receiving ARV treatment. One of the largest is the Sudan, have higher rates than those countries on the north- cost of the drugs. Until recently, all the major ARV drugs ern coast. were prohibitively expensive in developing countries. This Within Sub-Saharan Africa, the AIDS epidemic was situation is now changing, but ARV drugs and medical care noticed first in central Africa (Clumeck et al. 1983). Soon are still out of reach for the majority of AIDS patients in after, the epidemic was observed in East Africa, and sub- Africa. The global burden of AIDS is disproportionate in sequently in West Africa (Essex and Mboup 2002). The 238 | Souleymane Mboup, Rosemary Musonda, Fred Mhalu, and Max Essex epidemic seemed to occur last in southern Africa, although The overall burden that HIV/AIDS has placed on Sub- rates there are now the highest in Africa and in the world. Saharan Africa is unprecedented. It is now the most Six countries of southern Africa have adult prevalence rates common cause of death in the region (WHO 1999). It has of 20 percent or higher, and the mean prevalence rate for all been estimated that by the year 2010 life expectancy at birth of southern Africa is about 18 percent. could be decreased by at least 15 years in most of the region The mean adult prevalence rate in Sub-Saharan Africa is and by 30 years or more in five countries (Stanecki and 7.2 percent, whereas the mean rates in Asia are 0.4 percent Walker 2002). Population growth rates would fall and (see table 17.1). Haiti, with a prevalence rate of 5.6 percent, infant mortality rates would increase. It has been estimated has the highest outside of Africa, and Cambodia, with that mortality rates for children under five could experi- 2.6 percent, has the highest in Asia (UNAIDS 2004). Brazil, ence a fivefold increase in Botswana and Zimbabwe one of the few countries outside of the West that began ARV (Stanecki and Walker 2002). All these projections, however, treatment at an early stage, has an estimated adult preva- assume that ARV drugs are not being used to save lives in lence rate of 0.7 percent. the region. This is an assumption that, fortunately, is no longer valid. Table 17.1 HIV-1 Prevalence in Representative Regions and Countries RETROVIRUSES Regions and countries Prevalence (%) Dominant subtype East Africa 6.7 A (50%) Prior to the AIDS epidemic and the discovery of HIV, retro- Tanzania 8.8 Various viruses were known to exist and were associated with certain Uganda 4.1 A (60%) leukemias of animals and people. Such "oncoretroviruses" North Africa 0.3 B ( 95%) were characterized by an ability to replicate without killing Egypt, Rep. of 0.1 B ( 95%) the cells they infect. Lentiretroviruses, which include the Morocco 0.1 B ( 95%) HIVs, usually replicate at higher levels and kill the cells they Southern Africa 18.0 C ( 95%) infect. All retroviruses use reverse transcription of the viri- Botswana 37.3 C ( 95%) on RNA to replicate, and they produce a proviral DNA form Mozambique 12.2 C ( 95%) that can remain latent in host cell chromosomes. All retro- South Africa 21.5 C ( 95%) viruses also generate a high rate of mutations when tran- West Africa 4.4 CRF 02 (70%) scribing the RNA, but the error rate is highest for those Côte d'Ivoire 7.0 CRF 02 (80%) viruses that replicate to high levels, such as HIV. Nigeria 5.4 CRF 02 (70%) The high mutation rate facilitates a rapid rate of evolu- Senegal 0.8 CRF 02 (60%) tionary change for HIVs. Another feature of the viruses that Asia 0.4 Various promotes rapid variation is the diploid nature of the viral Cambodia 2.6 CRF 01 ( 90%) RNA. With two complete copies of the genome packaged China 0.1 CRF 07,08 (80%) in each virus particle, recombinational events are also fre- India 0.9 C (80%) quent, allowing progeny viruses to pick up large segments Thailand 1.5 CRF 01 ( 90%) of somewhat different genetic information in a short period Caribbean 1.6 B (90%) of time. Haiti 5.6 B (90%) An extremely important aspect of the rate of evolution- Latin America 0.6 B (90%) ary variation for HIVs is the selection pressure exerted by Brazil 0.7 B (80%) the host. This aspect is vividly illustrated by the rapid emer- North America 0.6 B ( 90%) gence of drug-resistant variations of HIV. Resistance occurs United States 0.6 B ( 90%) rapidly in a single individual, particularly when only one Western Europe 0.3 B ( 90%) drug is used. The same type of competition occurs during Italy 0.5 B (70%) host-mediated immunoselection pressure, in which viral Source: Country-specific prevalence data from UNAIDS 2004. Regional prevalence data from variants emerge to avoid control by epitope-specific host UNAIDS and WHO 2005 or adapted from UNAIDS 2004. Dominant subtype information from immune responses. The process is repeated frequently and Essex and Mboup 2002. Note: CRF circulating recombinant form. cyclically; the dominant clone of the virus elicits a new HIV/AIDS | 239 immune response, becomes controlled, and a new muta- subtypes, and viruses of one subtype differ from those of tional variant takes over as the next dominant viral clone. another subtype by an average of 20 to 30 percent in genetic Many of the mutations and recombinational events result sequence. Two of the six common categories of HIV-1s in the emergence of genomes that are incapable of repli- were found to be circulating recombinant forms (CRFs) cation. This occurs because viruses, to be viable, must fold (McCutchan 2000). their proteins in specific ways and retain reactive sites that The first major category of HIV-1 to be identified and signal key events. All HIVs, for example, must retain the categorized was subtype HIV-1B, which is responsible for ability to bind to the CD4 receptors and chemokine core- the epidemics in the Americas and Western Europe. These ceptors, such as CCR5 and CXCR4, on T lymphocytes and epidemics have been characterized by patterns that reflect macrophages. This in turn forces a certain level of conver- primary transmission via homosexual contact and injec- gent evolution. tion drug use. HIV-1B has also been found in other sites, As progressive cycles of HIV mutant clones begin to such as India, South Africa, and Thailand. Even in these exhaust the immune control capacity of the host, the num- sites, HIV-1B infections have not been associated with het- ber of dying lymphocytes exceeds the capacity for cell erosexual epidemics, although heterosexual epidemics due replacement and the number of CD4 lymphocytes rapidly to other HIV subtypes have occurred concurrently in the falls. As plasma viral load reaches levels of 50,000 copies per same sites (Hudgens et al. 2002; Janssens, Buve, and milliliter and above, and the number of CD4 lymphocytes Nkengasong 1997). drops to 300 per cubic millimeter and below, the ability of The most common HIV in the world and in Africa is the body to resist opportunistic infections is largely lost, and HIV-1 subtype C (HIV-1C). HIV-1C accounts for as many clinical AIDS develops. infections as all other HIVs combined, both in the world and in Africa (Essex 1999). It is responsible for the massive epidemic in southern Africa, and all the countries in Africa REGIONAL VARIATION that have the highest rates of HIV have HIV-1C epidemics. Along with an epidemic expansion that plateaued at the Soon after HIV-1 was identified, other lentiretroviruses were highest levels, the HIV-1C epidemic in southern Africa identified in primates. These included a wide range of simi- expanded faster than other HIV epidemics, although it an immunodeficiency viruses (SIVs) and a second category started later (Essex and Mboup 2002). of HIVs in people in West Africa, designated HIV-2s (Barin The next most important HIV virus in Africa, and in the et al. 1985). The SIVs naturally infected large fractions of world, is CRF 02, a recombinant form that originated from populations of various species of African monkeys, appar- an HIV-1A and an HIV-1G. This virus accounts for about ently without causing disease (Kanki, Homma, et al. 1985). 25 percent of the world's infections (Essex and Mboup The SIVs did, however, cause clinical AIDS when artificially 2002). It is largely responsible for the epidemics in West injected into Asian macaques (Daniel et al. 1985; Kanki, Africa and central Africa. Although mean prevalence rates in McLane, et al. 1985). these regions are only 5 percent and 6 percent, respectively, Some of the SIVs were virtually indistinguishable from they represent some sites with large populations, such as HIV-2s at the genetic level, although SIVs and HIV-2s were Nigeria. East Africa has had epidemics of HIV-1A and usually 50 to 60 percent different from HIV-1s. HIV-2s also HIV-1D, often coexisting in the same populations (Rayfield cause an AIDS-like disease in people, but with considerably et al. 1998). A few countries, such as Tanzania in the east and less efficiency than HIV-1s (Marlink et al. 1994). HIV-2s Cameroon in the west, have a wide variety of viruses present were much less transmissible, both by sexual contact and by in the same site (Fonjungo et al. 2000; Renjifo et al. 1999). In mother-to-infant transmission (Kanki, Sankale, and Mboup Tanzania, for example, an earlier epidemic of HIV-1A and 2002). Presumably because of this dramatic reduction in HIV-1D fused with the northward expansion of the HIV-1C transmission efficiency, HIV-2 has been largely limited to epidemic. This apparently resulted in the generation of a West Africa, and it has not caused any of the large epidemics large number of recombinant viruses (Renjifo et al. 1999). A typical of HIV-1. schematic distribution of the major subtypes and CRFs in Many different categories of HIV-1s have been identified, Africa is illustrated in figure 17.3. with at least six of them linked to major epidemics. The dif- Phenotypic distinctions between the HIV-1 subtypes and ferent categories of HIV-1s are usually designated clades, or CRFs are much less apparent than the differences between 240 | Souleymane Mboup, Rosemary Musonda, Fred Mhalu, and Max Essex Figure 17.3 Distribution of Major HIV-1 Subtypes PROGRESSION OF INFECTION and Circulating Recombinant Forms in Africa After initial exposure to HIV occurs, several weeks go by 30° 20° 10° 0° 10° 20° 30° 40° 50° IBRD 34534 MARCH 2006 North before the virus can be detected in blood. At least in the 0.3 case of HIV-1B infections in homosexual men, an acute, 30° 30° influenza-like illness then occurs in the majority of the Red infected. This is characterized by high levels of virus 20° Sea 20° replication reflected in the blood as viremia. Whether the same flu-like illness occurs at the same rate in heterosexual 10° 10° Western infections of other HIV-1 subtypes is unclear. 4.4 Horn During the stage of acute viremia large numbers of 4.3 T lymphocytes become infected in lymph nodes (Pantaleo 0° ATLANTIC 0° OCEAN Central 5.3 Eastern et al. 1993), and patients are highly infectious to other 6.7 10° 10° potential contacts (Quinn and Chaisson 2003). The acute viremia then falls precipitously, presumably because of an effective immune response, albeit a response that later 20° B 20° A/G becomes largely ineffective as mutant variants emerge. The A, D, and C C Southern magnitude of the acute viremia and the depth of the subse- no data INDIAN 30° 18.0 30° international boundaries OCEAN quent resolution, sometimes called the set point, probably determine the subsequent rate of disease progression. For a This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° minority of those infected, a rapid and well-controlled set Source: Essex and Mboup 2002; UNAIDS 2004; UNAIDS and WHO 2005. point apparently results in the development of clinical AIDS only after a very prolonged period of time. Such individuals are sometimes called "long-term nonprogressors." HIV-1s and HIV-2s. HIV-1C genomes ordinarily have The length of time before disease development may also three NFKB enhancer sequences, whereas other subtypes be related to other factors, such as the infected individual's have only two, and in the case of HIV-2s, only one (Montano genetic background (Winkler and O'Brien 2002), nutri- et al. 1997). This is almost certainly associated with the tional state, and parasite burden of coinfecting microorgan- higher transcriptional activation rates seen for HIV-1Cs isms. Antigenic stimulation of infected cells causes DNA (Montano et al. 2000). It is also a possible explanation for synthesis, which in turn causes activation and replication of the higher rates of genomic variation (Novitsky et al. 1999) latent HIV. The tissue damage caused by other infections and the higher viral loads reported for HIV-1Cs (Neilson also stimulates the release of inflammatory cytokines, which et al. 1999). The efficient spread of HIV-1C is also compati- also causes transcriptional activation of HIV (Montano ble with this subtype's preference for the CCR5 coreceptor et al. 2000). With all these variables, however, clinical AIDS (Ping et al. 1999; Tscherning et al. 1998). HIV-1Ds and rarely occurs before 4 to 5 years after infection, and almost HIV-1Bs, for example, are much more likely to use the always occurs within 10 to 12 years after infection. In the CXCR4 coreceptor than the HIV-1Cs (Tscherning et al. absence of therapy, death usually occurs within 1 to 3 years 1998), and neither HIV-1B nor HIV-1D has spread to cause after the onset of clinical AIDS. large epidemics in Africa. Coinfection of individuals with two different HIVs-- different subtypes; different types, such as HIV-1 and PREVENTION HIV-2; or different variants within a subtype--can occur. These events appear to occur less often than might be Methods of preventing HIV infection can be divided into expected (Travers et al. 1995), perhaps because of some level those that are currently available, such as education, and of cross-protection due to activation of chemokines, recep- those that are not yet available but are being pursued tor competition, or specific immunity. When the same cells through research, such as vaccines. are coinfected, the opportunity for generation of infectious HIV transmission through blood transfusion, although recombinants is substantial. an important issue at the early stages of the pandemic, now HIV/AIDS | 241 occurs only rarely. The serology tests used to determine if that are used for treatment of AIDS, are highly effective if blood is contaminated are highly sensitive and specific, given within hours after the presumed exposure. The same although an infectious unit of blood can occasionally be interventions can be used for victims of sexual violence. missed if the donor was infected within the two to three weeks before donation, before antibodies had time to develop. The use of contaminated needles or injection VACCINES AND MICROBICIDES equipment is sometimes an important method of transmis- sion in defined populations, but it is not nearly as important Effective vaccines that exist against other viral diseases are as sexual transmission, at least in Africa. based on the injection of either killed virus, purified viral Voluntary testing and counseling programs are extremely surface proteins, or live attenuated virus. All three types are important to reduce sexual transmission. A major limitation commonly used in people. The Salk polio vaccine is killed in most countries is the reluctance of most people to get virus, the vaccine used against hepatitis B is a purified viral tested and thus learn their status. Such reluctance may be surface protein, and the Sabin oral polio vaccine is a live particularly evident where infected individuals have few or attenuated virus. The killed virus and viral protein no opportunities for treatment with ARV. In such situations approaches can function only by inducing virus-neutralizing concerns about stigma and death from the infection may antibodies. The live attenuated types of vaccine often induce provide disincentives for the individual to learn his or her both neutralizing antibodies and cytolytic T cell immunity. status. The use of live attenuated HIV as a vaccine was largely Condoms are highly effective when used properly. dismissed, even at the earliest stages, because of safety con- Abstinence provides a guarantee against sexual infection. cerns. Killed HIV was similarly dismissed for safety concerns But neither abstinence nor condoms allow couples to have and because it was assumed that subunit surface proteins children. would work as well as killed virus without the same safety The use of ARVs for chemoprophylaxis is effective in pre- concerns, as was true of the hepatitis B vaccine. As a result, venting transmission of HIV from infected mothers to their almost all the initial experimental vaccines were based on infants. In the absence of intervention, 25 to 45 percent of the use of HIV proteins gp120 or gp160 that protrude from infants born to HIV-positive mothers become infected. The the outer surface of the virus. infection occurs in utero, during the process of birth, and As HIVs grown in cultured cell lines provided the highest through breastfeeding. The use of ARVs, such as zidovudine titers for subsequent purification, the first vaccines were (AZT), was shown to reduce neonatal transmission by as made from T-cell-line adapted gp120/160 proteins. It was much as 67 percent in nonbreastfeeding populations if soon recognized that the neutralizing antibodies these vac- given at least six to eight weeks before birth (Connor et al. cines induced were ineffective against naturally occurring 1994). Even when given only at the time of labor, nevirapine HIVs. Subsequently it was also recognized that monomeric (NVP) apparently reduced intrapartum transmissions by gp120/160 proteins induced antibodies that would work as much as 50 percent (Guay et al. 1999). The use of drug only against the exact strain selected. The profound combinations early in gestation can presumably reduce in sequence variation seen among HIVs created a situation in utero and intrapartum transmission to only a few percent. which matching the vaccine gp120 to the field challenge Avoidance of breastfeeding can obviously eliminate post- strain was not possible. natal infections by this route. However, in many cultures The next wave of HIV vaccine research was based prima- recommendations for formula feeding are not well received, rily on the development of cytolytic T cell (CTL) vaccines. in part because of the stigma associated with bottle feeding Most were based on the use of nonvirulent viruses, such as of the infant. The effectiveness of chemoprophylaxis to the vaccinia, the vaccine used for smallpox, or adenoviruses. infant or the mother, or both, while breastfeeding is being Genes, or parts of genes, of HIV that were suspected of being evaluated. able to induce CTL responses were inserted in the "vaccine ARV drugs can also be used to block transmission of HIV vector viruses" through gene splicing. HIV antigens or "CTL by accidental needlestick infections, such as those that might epitopes" must be delivered in this way, because for immu- occur among medical personnel treating AIDS patients nity to be effective, recognition of the immunogen must be (Bouvet, Laporte, and Tarantola 2002). The same three-drug in conjunction with cell processing to match the HIV anti- combinations, such as AZT, lamivudine (3TC), and NVP, gens to the histocompatibility antigens of the recipient. 242 | Souleymane Mboup, Rosemary Musonda, Fred Mhalu, and Max Essex Several CTL vaccines have now been evaluated in people which ranged from gastrointestinal problems and anemias for safety and immunogenicity. In general, they appear to be to peripheral neuropathies. The other was the rapid genera- quite safe, but they do not appear to be sufficiently stimu- tion of drug-resistant variants of HIV, which soon grew just lating to elicit strong immune responses that are likely to be as well in the patient as before the drugs were first used. protective. Developing CTL vaccines using other viruses as Using two drugs at once lowered HIV viremia even better vectors is an entirely new area of vaccine science. If these than using one drug and delayed the time to development of methods can be mastered, they could probably also be drug resistance. However, even with two NRTIs, drug resist- extrapolated for use in combating many other diseases. ance often developed within a year, and resistance to one of A major part of the rationale for CTL vaccines was the the NRTIs often prompted resistance to other NRTIs of the assumption that they would be more broadly cross-reactive same class. than gp120 vaccines. Viral core proteins are usually more The next class of drugs available was the nonnucleoside genetically conserved than viral surface proteins, thus reverse transcriptase inhibitors (NNRTIs), such as nevirap- reducing antigenic variation and reducing opportunities for ine (NVP) and efavirenz (EFV). These drugs were even eas- immune selection in the vaccinee to evade effectiveness. ier for the virus to mutate around to cause drug resistance, Even for CTL vaccines, however, it seems prudent to use the and this happened within weeks when one of these drugs relevant HIV-1 subtype or regional strain to increase the was given alone, as "monotherapy." However, when given likelihood of protection. with two or more other drugs, they could often be given for To maximize the chances that a vaccine may work, most long periods of time. researchers endorse the use of vaccines or vaccine combina- The third major class of drugs, the protease inhibitors, tions that might induce both neutralizing antibodies and work on an entirely different gene of HIV, so they provided CTL responses. This is ordinarily done using a"prime-boost" an additional, separate mechanism to keep down virus repli- strategy of several inoculations, in which the first (prime) cation. However, unlike the NRTIs and the NNRTIs, which would allow the induction of CTL responses, and the boost could often be made "off patent" or as "generics," the pro- would encourage both CTL and antibody responses. tease inhibitors are still expensive and generally not available Recently, research to design gp120 antigens that would in developing countries. induce cross-reactive neutralizing antibodies has also begun The standard drug regimen for most developing coun- a new generation of vaccine approaches (Fouts et al. 2000). tries is now a three-drug combination that includes two The concept is based on the stabilization of a conformation- NRTIs, especially AZT and 3TC, and one NNRTI, such as al state of the HIV-1 gp120 at the time it interacts with the NVP. In most cases this regimen works well, unless resist- CD4 receptor or the CCR5 receptor or both. This and other ance to any of the drugs is already present. This happens, for approaches serve to illustrate that HIV vaccine research is example, if women have recently been treated with AZT or alive and well in providing new designs, but is also unlikely NVP alone, as part of a chemoprevention strategy to block to yield a final product of high efficacy in less than 5 to infection of their infant. The use of NVP alone during labor, 10 years. Thus, the prevention of further spread of HIV for example, can cause resistance to the whole NNRTI class within the next decade must be centered on other measures. of drugs in 20 percent or more of mothers, even when only one dose is given (Eshleman et al. 2001). In the West, combination three-drug therapy usually THERAPY begins earlier than the current recommendations projected for the developing world. In the United States, for example, The first phase of therapy using ARV drugs was AZT. drug therapy would usually be recommended by the time Initially it was the only drug, and when used alone it often the patient fell below 350 CD4 cells per cubic millimeter or gave AIDS patients another six months or so in partial a viral load of less than 25,000 RNA per milliliter. In Africa remission. Soon, other related drugs, nucleoside analogue a consensus is building that the most logical time to initiate reverse transcriptase inhibitors (NRTIs) such as lamivudine therapy is when CD4 cells fall below 200, or when the (3TC), didanosine (DDI), and stavudine (D4T), were also patient has experienced an AIDS-defining illness. By this available. One of these drugs used with AZT clearly worked time, most patients have viral loads in excess of 50,000 RNA better than using just one drug alone. However, their use per milliliter. In early studies, disease-free one-year survival posed two significant problems. One was drug toxicities, rates for patients given a three-drug combination with HIV/AIDS | 243 drugs that cost more, as do monitoring costs and the initiation below 200 CD4 cells per cubic millimeter seem involvement of personnel with more specialized skills. The to be quite good (Djomand et al. 2003). Rigorous adherence use of CD4 counts alone may be adequate for initiation, for to taking the drugs is essential to increase the time to devel- example, but viral load tests and drug resistance tests may be opment of drug-resistant variants. Many newer drugs and necessary if drug failure occurs. The latter tests are more drug combinations reduce the difficulty of dosing because expensive, but they may be needed to validate the efficacy of they are taken only once per day and have fewer gastroin- second- or third-line drug regimens. testinal side effects. Studies using several combinations of three drugs to treat AIDS patients have now been conducted in several African countries (Coetzee et al. 2004; Djomand et al. 2003; Laurent CONCLUSION et al. 2002; Weidle et al. 2002; Wester et al. 2005). The results The epidemic of HIV/AIDS is unprecedented, having indicate a high degree of success. Despite the initiation of expanded from a new disease to the leading cause of death therapy at low levels of immune competence (for example, in Sub-Saharan Africa in just over two decades. Despite 10­200 CD4 cells per cubic millimeter), responses have some clear examples of success, such as Senegal for preven- been good (Coetzee et al. 2004; Laurent et al. 2002; Weidle tion, and Uganda for control, major new measures are et al. 2002; Wester et al. 2005). These include survival rates needed to avoid further devastation. The epidemic in Africa of 85 percent or higher, and escalation of CD4 cell counts is uneven, with the greatest burden in southern Africa, of 150 CD4 cells per cubic millimeter or more by a year where populations are already experiencing major reduc- after drug initiation. Even patients who began drug therapy tions in life expectancy and a reversal of progress in the at cell counts below 50 CD4 cells per cubic millimeter, management of national economies. most of whom would probably die within a year if left About two-thirds of the world's HIV infections are in untreated, had surprisingly positive responses. Sub-Saharan Africa in just 10 percent of the world's popula- The positive responses occurred in part because most tion. Because HIV/AIDS in the Americas and Europe is patients showed a strong commitment to adhere to the pre- uncommon in women and rare in children, Africa holds scribed regimen of drugs. Although different viral subtypes more than 90 percent of the world's burden on such issues may show different profiles of drug mutations (Quan et al. as child mortality and the care of orphans. 2003; Turner et al. 2004), high levels of adherence reduced At present, prevention strategies are largely limited to rates of clinical failure due to drug resistance (Coetzee et al. education for changes in sexual practices and the use of con- 2004; Laurent et al. 2002; Wester et al. 2005). It is too early doms. Vaccines and microbicides represent an extremely to know whether HIV drug resistance will become a major important area of research, but useful products are not likely problem in Africa. However, the widespread use of just one to be available in less than 5 to 10 years. ARV drug treatment or two drugs in perinatal chemoprophylaxis clearly results programs have begun in Africa but are as yet very modest in in high levels of genotypic resistance in mothers who may impact. Operational research to maximize the benefits of soon need the same drugs for their own therapy. Whether ARV use will be important to address treatment efficacy and such drug-resistant variants will be transmitted by sexual drug resistance, the impact on health manpower resources, contact remains to be determined. and cost-effectiveness. The full devastation of AIDS in The cost of a basic three-drug regimen, such as AZT, Africa has yet to be fully appreciated. Still, an increase in 3TC, and NVP, is now as low as about US$1 per day. awareness and willingness to act should give cause for Although this is far below the original cost for the same optimism. drugs, it is still above a level that is possible for many in the poorest countries. The actual cost of treatment, including costs of health care personnel, infrastructure, and labora- REFERENCES tory diagnosis and monitoring, is generally substantially higher than the cost of the drugs alone. Initiation on ARV Barin, F., S. Mboup, F. Denis, P. Kanki, J. S. Allan, T. H. Lee, and therapy should be possible in many sites in Africa, but costs M. Essex. 1985. "Serological Evidence for Virus Related to Simian T-Lymphotropic Retrovirus III in Residents of West Africa." Lancet 2: will also increase as patients become resistant to any of those 1387­89. drugs used in the "first-line" regimen. Switching to different Barre-Sinoussi, F., J. C. Chermann, F. Rey, M. T. Nugeyre, S. Chamaret, regimens may require the use of protease inhibitors or other J. Gruest, C. Dauguet, et al. 1983. "Isolation of a T-Lymphotropic 244 | Souleymane Mboup, Rosemary Musonda, Fred Mhalu, and Max Essex Retrovirus from a Patient at Risk for Acquired Immune Deficiency Janssens, W., A. Buve, and J. N. Nkengasong. 1997. "The Puzzle of HIV-1 Syndrome (AIDS)." Science 220: 868­71. Subtypes in Africa." AIDS 11: 705­12. Bouvet, E., A. Laporte, and A. Tarantola. 2002. "Postexposure Prophylaxis Kanki, P., J. L. Sankale, and S. Mboup. 2002. "Biology of Human for Occupational Exposure and Sexual Assault." In AIDS in Africa, ed. Immunodeficiency Virus Type 2." In AIDS in Africa, ed. M. Essex, M. Essex, S. Mboup, P. Kanki, R. Marlink, and S. Tlou, 571­83. New S. Mboup, P. Kanki, R. Marlink, and S. Tlou, 74­103. New York: Kluwer York: Kluwer Academic/Plenum Publishers. Academic/Plenum Publishers. CIA (Central Intelligence Agency). 2006. The World Factbook. http://www. Kanki, P. J., T. Homma, T. H. Lee, N. W. King Jr., R. D. Hunt, and M. Essex. cia.gov/cia/publications/factbook/ (accessed January 13, 2006). 1985. "Antibodies to Human T-Cell Leukemia Virus-Membrane Clumeck, N., F. Mascart-Lemone, J. de Maubeuge, D. Brenez, and Antigens in Macaques with Malignant Lymphoma." Haematology and L. Marcelis. 1983. "Acquired Immune Deficiency Syndrome in Black Blood Transfusion 29: 345­49. Africans." Lancet 1: 642. Kanki, P. J., M. F. McLane, N. W. King Jr., N. L. Letvin, R. D. Hunt, P. Sehgal, Coetzee, D., K. Hildebrand, A. Boulle, G. Maartens, F. Louis,, V. Labatala, M. D. Daniel, R. C. Desrosiers, and M. Essex. 1985. "Serologic H. Reuter, N. Ntwana, and E. Goemaere. 2004. "Outcomes after Two Identification and Characterization of a Macaque T-Lymphotropic Years of Providing Antiretroviral Treatment in Khayelitsha, South Retrovirus Closely Related to HTLV-III." Science 228: 1199­1201. Africa." AIDS 18: 887­95. Kitchen, L. W., F. Barin, J. L. Sullivan, M. F. McLane, D. B. Brettler, P. H. Connor, E. M., R. S. Sperling, R. Gelber, P. Kiselev, G. Scott, M. J. Levine, and M. Essex. 1984. "Aetiology of AIDS--Antibodies to O'Sullivan, R. VanDyke, et al. 1994. "Reduction of Maternal-Infant Human T-Cell Leukaemia Virus (type III) in Haemophiliacs." Nature Transmission of Human Immunodeficiency Virus Type 1 with 312: 367­69. Zidovudine Treatment." Pediatric AIDS Clinical Trials Group Protocol Laurent, C., N. Diakhaté, N. F. N. Gueye, M. A. Touré, P. S. Sow, M. A. Faye, 076 Study Group. New England Journal of Medicine 331: 1173­80. M. Gueye, et al. 2002. "The Senegalese Government's Highly Active Daniel, M. D., N. L. Letvin, N. W. King, M. Kannagi, P. K. Sehgal, R. D. Antiretroviral Therapy Initiative: An 18-Month Follow-Up Study." Hunt, P. J. Kanki, M. Essex, and R. C. Desrosiers. 1985. "Isolation of AIDS 16: 1363­70. T-Cell Tropic HTLV-III-Like Retrovirus from Macaques." Science 228: Marlink, R., P. Kanki, I. Thior, K. Travers, G. Eisen, T. Siby, I. Traore, C. C. 1201­4. Hsieh, M. C. Dia, and E. H. Gueye. 1994. "Reduced Rate of Disease Djomand, G., T. Roels, T. Ellerbrock, D. Hanson, F. Diomande, B. Monga, Development after HIV-2 Infection as Compared to HIV-1." Science C. Maurice, et al. 2003. "Virologic and Immunologic Outcomes and 265: 1587­90. Programmatic Challenges of an Antiretroviral Treatment Pilot Project McCutchan, F. E. 2000. "Understanding the Genetic Diversity of HIV-1." in Abidjan, Côte d'Ivoire." AIDS 17: S5­S15. AIDS 14 (Suppl. 3): S31­44. Duesberg, P. 1988. "HIV Is Not the Cause of AIDS." Science 241: 514-17. Montano, M. A., C. P. Nixon, T. Ndung'u, H. Bussmann, V. A. Novitsky, Eshleman, S. H., M. Mracna, L. A. Guay, M. Deseyve, S. Cunningham, D. Dickman, and M. Essex. 2000. "Elevated Tumor Necrosis Factor- M. Mirochnick, P. Musoke, et al. 2001. "Selection and Fading of Alpha Activation of Human Immunodeficiency Virus Type 1 Resistance Mutations in Women and Infants Receiving Nevirapine to Subtype C in Southern Africa Is Associated with an NF-Kappab Prevent HIV-1 Vertical Transmission (HIVNET 012)." AIDS 15: Enhancer Gain-of-Function." Journal of Infectious Diseases 181: 76­81. 1951­57. Montano, M. A., V. Novitsky, J. Blackard, N. Cho, D. Katzenstein, and Essex, M. 1999. "Human Immunodeficiency Viruses in the Developing M. Essex. 1997. "Divergent Transcriptional Regulation among World." Advances in Virus Research 53: 71­88. Expanding Human Immunodeficiency Virus Type 1 Subtypes." Journal of Virology 71: 8657­65. Essex, M., and S. Mboup. 2002. "Regional Variations in the African Epidemics." In AIDS in Africa, ed. M. Essex, S. Mboup, P. Kanki, R. Neilson, J. R., G. C. John, J. K. Carr, P. Lewis, J. K. Kreiss, S. Jackson, R. W. Marlink, and S. Tlou, 631­40. New York: Kluwer Academic/Plenum Nduati, et al. 1999. "Subtypes of Human Immunodeficiency Virus Publishers. Type 1 and Disease Stage among Women in Nairobi, Kenya." Journal of Virology 73: 4393­4403. Fonjungo, P. N., E. N. Mpoudi, J. N. Torimiro, G. A. Alemnji, L. T. Eno, J. N. Nkengasong, F. Gao, et al. 2000. "Presence of Diverse Human Novitsky, V., M. Montano, M. McLane, B. Renjifo, F. Vannberg, B. Foley, Immunodeficiency Virus Type 1 Viral Variants in Cameroon." AIDS T. Ndung'u, et al. 1999. "Molecular Cloning and Phylogenetic Analysis Research and Human Retroviruses 16: 1319­24. of Human Immunodeficiency Virus Type 1 Subtype C: A Set of 23 Full-Length Clones from Botswana." Journal of Virology 73: 4427­32. Fouts, T. R., R. Tuskan, K. Godfrey, M. Reitz, D. Hone, G. K. Lewis, and A. L. DeVico. 2000. "Expression and Characterization of a Single-Chain Pantaleo, G., C. Graziosi, J. F. Demarest, L. Butini, M. Montroni, C. H. Fox, Polypeptide Analogue of the Human Immunodeficiency Virus Type 1 J. M. Orenstein, D. P. Kotler, and A. S. Fauci. 1993. "HIV Infection Is Gp120-CD4 Receptor Complex." Journal of Virology 74, 11427­36. Active and Progressive in Lymphoid Tissue during the Clinically Latent Stage of Disease." Nature 362: 355­58. Gallo, R. C., S. Z. Salahuddin, M. Popovic, G. M. Shearer, M. Kaplan, B. F. Haynes, T. J. Palker, et al. 1984. "Frequent Detection and Isolation of Ping, L., J. Nelson, I. Hoffman, J. Schock, S. Lamers, M. Goodman, Cytopathic Retroviruses (HTLV-III) from Patients with AIDS and at P. Vernazza, et al. 1999. "Characterization of V3 Sequence Risk for AIDS." Science 224: 500­03. Heterogeneity in Subtype C Human Immunodeficiency Virus Type 1 Isolates from Malawi: Underrepresentation of X4 Variants." Journal of Guay, L. A., P. Musoke, T. Fleming, D. Bagenda, M. Allen, C. Nakabiito, J. Virology 73: 6271­81. Sherman, et al. 1999. "Intrapartum and Neonatal Single-Dose Nevirapine Compared with Zidovudine for Prevention of Mother-to- Popovic, M., M. G. Sarngadharan, E. Read, and R. C. Gallo. 1984. Child Transmission of HIV-1 in Kampala, Uganda: HIVNET 012 "Detection, Isolation, and Continuous Production of Cytopathic Randomised Trial." Lancet 354: 795­802. Retroviruses (HTLV-III) from Patients with AIDS and Pre-AIDS." Science 224: 497­500. Hudgens, M. G., I. M. Longini Jr., S. Vanichseni, D. J. Hu, D. Kitayaporn, P. A. Mock, M. E. Halloran, G. A. Satten, K. Choopanya, and T. D. Quan, Y., B. Brenner, R. Marlink, M. Essex, T. Kurimura, and M. Wainberg. Mastro. 2002."Subtype-Specific Transmission Probabilities for Human 2003."Drug Resistance Profiles of Recombinant Reverse Transcriptases Immunodeficiency Virus Type 1 among Injecting Drug Users in from HIV-1 Subtypes A/E, B, and C." AIDS Research and Human Bangkok, Thailand." American Journal of Epidemiology 155: 159­68. Retroviruses 19: 743­53. HIV/AIDS | 245 Quinn, T. C., and R. E. Chaisson. 2003. "International Epidemiology of and Amino Acid Polymorphisms at Drug Resistance Sites in Non- Human Immunodeficiency Virus." In Infectious Diseases, ed. S. L. B-Subtype Variants of Human Immunodeficiency Virus Type 1." Gorback, J. G. Bartlett, and N. R. Blacklow, 970­86. Philadelphia: Antimicrobial Agents and Chemotherapy 48: 2993­98. Lippincott Williams & Wilkins. UNAIDS (Joint United Nations Programme on HIV/AIDS). 2004. Rayfield, M. A., R. G. Downing, J. Baggs, D. J. Hu, D. Pieniazek, C. C. Luo, Report on the Global HIV/AIDS Epidemic. Geneva: UNAIDS and B. Biryahwaho, R. A. Otten, S. D. Sempala, and T. J. Dondero. 1998. WHO. "A Molecular Epidemiologic Survey of HIV in Uganda." HIV Variant UNAIDS and WHO (World Health Organization). 2005. AIDS Epidemic Working Group. AIDS 12: 521­27. Update: December 2005. Geneva: UNAIDS and WHO. Renjifo, B., P. Gilbert, B. Chaplin, F. Vannberg, D. Mwakagile, G. Weidle, P. J., S. Malamba, R. Mwebaze, C. Sozi, G. Rukundo, R. Downing, Msamanga, D. Hunter, W. Fawzi, and M. Essex. 1999. "Emerging D. Hanson, et al. 2002. "Assessment of a Pilot Antiretroviral Drug Recombinant Human Immunodeficiency Viruses: Uneven Therapy Programme in Uganda: Patients' Response, Survival, and Representation of the Envelope V3 Region." AIDS 13: 1613­21. Drug Resistance." Lancet 360: 34­40. Stanecki, K. A., and N. Walker. 2002. "Current Estimates and Projections Wester, C. W., S. Kim, H. Bussmann, A. Avalos, N. Ndwapi, T. F. Peter, for the Epidemic." In AIDS in Africa, ed. M. Essex, S. Mboup, P. Kanki, T. Gaolathe, et al. 2005. "Initial Response to Highly Active R. Marlink, and S. Tlou, 281­96. New York: Kluwer Academic/Plenum Antiretroviral Therapy in HIV-1C Infected Adults in a Public Sector Publishers. Treatment Program in Botswana." Journal of Acquired Immune Travers, K., S. Mboup, R. Marlink, A. Gueye-Ndiaye, T. Siby, I. Thior, Deficiency Syndrome 40: 336­43. Traore, et al. 1995. "Natural Protection against HIV-1 Infection WHO (World Health Organization). 1999. The World Health Report Provided by HIV-2." Science 268: 1612­15. 1999--Making a Difference. Geneva: WHO. Tscherning, C., A. Alaeus, R. Fredriksson, A. Bjorndal, H. Deng, D. R. ------. 2005. Progress on Global Access to HIV Antiretroviral Therapy : An Littman, E. M. Fenyo, and J. Albert. 1998. "Differences in Chemokine Update on "3 by 5". Geneva: World Health Organization. Coreceptor Usage between Genetic Subtypes of HIV-1." Virology 241: Winkler, C., and S. O'Brien. 2002. "Effect of Genetic Variation on HIV 181­88. Transmission and Progression to AIDS." In AIDS in Africa, ed. Turner, D., B. Brenner, D. Moisi, M. Detorio, R. Cesaire, T. Kurimura, M. Essex, S. Mboup, P. Kanki, R. Marlink, and S. Tlou. 52­73. New H. Mori, M. Essex, S. Maayan, and M. A. Wainberg. 2004. "Nucleotide York: Kluwer Academic/Plenum Publishers. 246 | Souleymane Mboup, Rosemary Musonda, Fred Mhalu, and Max Essex Chapter 18 Lifestyle and Related Risk Factors for Chronic Diseases Krisela Steyn and Albertino Damasceno Chronic diseases, often referred to as noncommunicable ANTENATAL INFLUENCES ON THE EMERGENCE diseases (NCDs), usually emerge in middle age after long OF RISK FACTORS FOR CHRONIC DISEASES exposure to an unhealthy lifestyle involving tobacco use, a lack of regular physical activity, and consumption of The fetal origins of adult chronic diseases play a particularly diets rich in highly saturated fats, sugars, and salt, typified important role in Sub-Saharan Africa countries. The ade- by "fast foods." This lifestyle results in higher levels of risk quacy of the mother's nutrition before and during preg- factors, such as hypertension, dyslipidemia, diabetes, and nancy is the first key component in determining the infant's obesity that act independently and synergistically. The birthweight. The latter in its own right is associated with the risk factors are frequently undiagnosed or inadequately emergence of chronic disease risk factors in these children managed in health services designed to treat acute (Barker 1993, 1994). In The Gambia, an inverse relationship conditions. was found between the weight gain of women in the last Chronic conditions are frequently incorrectly considered trimester of pregnancy and the blood pressure of their chil- to have limited impact on the burden of disease in Sub- dren at age eight years (Margetts et al. 1991). Saharan Africa, because of the known high relevance of Another critical factor during pregnancy is cigarette the infectious diseases. Nevertheless, these diseases occur smoking, which results in high rates of low birthweight in younger age groups more commonly in Sub-Saharan (LBW) babies and other complications in pregnancy. This Africa than in the developed countries and are at least as association has been reported in some Sub-Saharan Africa common in the poor sector of society as in the more countries (Odendaal, Van Schie, and De Jeu 2001; K. Steyn, affluent. unpublished data). Fortunately, smoking tobacco products The current burden of chronic diseases reflects the during pregnancy or the reproductive years is not a com- cumulative effects of unhealthy lifestyles and the resulting mon phenomenon in African women, particularly if they risk factors over the life span of people. Some of these influ- still live according to their traditional lifestyles. However, ences are present from before a child is born. surveys of women who progressively adopt Western 247 lifestyles show that smoking during pregnancy or during situation led to more starvation and more smoking during the reproductive years is becoming far more common pregnancy and thus to the birth of more LBW babies in (Steyn et al. 1994; Steyn et al. 1997). these countries. LBW is common in Sub-Saharan Africa countries. For example, Kinabo, Kissawke, and Msuya (1997) reviewed the birth records of 40,595 full-term live singleton infants TOBACCO USE and found a 15.4 percent prevalence of birthweight below 2,500 grams. In a five-year randomized control trial of As tobacco-control activities in the developed world have maternal supplementation in a rural primary health care increased, the tobacco industry has shifted its marketing to setting in The Gambia, 15.9 percent of LBW babies were middle- and low-income countries. Not only do these coun- born in the control group. The intervention group received tries have fewer formal tobacco-control activities in place, high-energy groundnut biscuits for about the last 20 weeks but they also have a much larger population, which can pro- of their pregnancy, and that significantly reduced the LBW vide future consumers of tobacco products. In Sub-Saharan rate to 10.7 percent. These data highlight the need to ensure Africa countries, most people are young and have relatively adequate nutrition for women in Sub-Saharan Africa during low levels of education. They rarely receive the necessary pregnancy (Ceesay et al. 1997). health education to allow them to critically evaluate the Levitt, Steyn, De Wet, and colleagues (1999) showed that material provided by the tobacco industry promotions. LBW predicted systolic blood pressure level at age five years The following five countries participated in the global in the children in Soweto, South Africa. The data show that Youth Tobacco Survey: Ghana, Malawi, Nigeria, South children who had a LBW and had gained the most weight Africa, and Zimbabwe (Global Youth Tobacco Survey during the five years since birth had the highest systolic Collaborative Group 2002). The data from the five Sub- blood pressure. Those children who were of normal weight Saharan Africa countries are shown in table 18.1. Although at birth and maintained a normal weight had the lowest sys- a significant number of youths age 13 to 15 smoked ciga- tolic blood pressure (Levitt, Steyn, De Wet, et al. 1999). A rettes, many more used other tobacco products. Relatively glucose tolerance test done on these children when they few of the youth smoked at home. These data also raise con- were seven years old showed that LBW in conjunction with cern about the pattern and impact of tobacco marketing in rapid childhood weight gain, especially if there was a large Sub-Saharan Africa. Overall, more than 10 percent of the gain of subcutaneous fat, produced poor glucose tolerance youth had been offered free cigarettes by the tobacco indus- (Crowther et al. 1998). try, and about 40 percent of them thought boys who smoked A study by Longo-Mbenza and colleagues (1999) in had more friends. Despite the traditional taboos against Kinshasa, Democratic Republic of Congo, of 2,409 school- smoking by women, about 20 percent thought that those children age 5 to 16 years also showed a significant inverse girls who smoked had more friends than those who didn't. relationship between LBW (less than 2,500 grams) and More than 60 percent of the youth had seen antismoking blood pressure, as well as heart rate. In Cape Town, South media messages, and between one-third and two-thirds had Africa, Levitt, Steyn, Lambert, and colleagues (1999) found been taught the dangers of smoking at school. Of the youth the same relationship in adults age 20 years who had LBW. who smoked, 74 percent and 54 percent in South Africa and These young adults also had impaired glucose tolerance, as Zimbabwe, respectively, had tried to quit in the previous well as increased plasma cortisol and cortisol axis activation, year (Global Youth Tobacco Survey Collaborative Group compared with young adults who had a birthweight of 2002). 2,500 grams or more (Levitt et al. 2000; Longo-Mbenza The impact of tobacco marketing on five-year-olds has et al. 1999). been shown in children living in Johannesburg and Soweto The association between poor intrauterine growth in the Birth-to-Ten study (Levitt, Steyn, De Wet, et al. 1999). because of inadequate nutrition and smoking tobacco This birth cohort comprised children who were born in during pregnancy, resulting in LBW and increased NCD risk 1990, and by the time they were five years old the new politi- in children in Sub-Saharan Africa, could possibly increase cal dispensation had been established and the country had a in the twenty-first century. At the beginning of this century new South African flag. When these children were shown most of the Sub-Saharan Africa countries were in the grip pictures of well-known logos of cigarettes available in South of serious food shortages as well as unopposed promotion Africa along with that of the new flag and the logos of other of tobacco products targeting women and youth. This well-known South African brands, over 90 percent of the 248 | Krisela Steyn and Albertino Damasceno Table 18.1 Exposure to Tobacco Products of Participants, Age 13­15 Years, in the Global Youth Tobacco Survey, 1999­2001 Site Blantyre, Lilongwe, Cross River South Harare, Maricaland, Students Ghana Malawi Malawi State, Nigeria Africa Zimbabwe Zimbabwe Number 1,088 783 1,083 914 2,579 621 700 Currently smoking 4.2 2.4 6.1 7.0 17.6 10.7 10.0 cigarettes (1.7) (2.2) (1.9) (3.0) (2.5) (3.4) (3.7) Currently using other 14.5 14.7 12.9 14.0 11.8 9.5 13.2 tobacco productsa (3.4) (2.8) (2.1) (3.2) (3.4) (3.4) (4.5) Current smokers who 24.4 n.a. 30.4 22.2 18.8 25.2 26.0 usually smoke at home (14.0) (12.2) (15.3) (4.2) (12.2) (10.7) Exposed to smoke from 22.2 19.0 16.0 34.3 43.6 36.2 35.0 others in their home (3.8) (4.5) (2.3) (5.1) (4.6) (5.0) (6.0) Think boys who smoke 41.1 41.6 48.8 42.5 48.1 43.1 43.8 have more friends (5.0) (4.9) (5.2) (4.6) (6.2) (5.8) (3.7) Think girls who smoke 30.1 21.1 20.3 26.9 30.7 23.3 18.6 have more friends (5.2) (4.2) (3.4) (4.0) (4.0) (3.9) (2.8) Think smoking should be 58.2 90.1 85.1 60.2 53.4 43.2 31.6 banned from public places (8.8) (3.0) (6.8) (4.6) (9.1) (11.1) (8.1) Think ETS is harmful 41.6 83.1 81.8 35.4 57.3 45.3 31.0 to themselves (9.7) (3.6) (6.1) (5.9) (7.5) (6.2) (6.3) Reported having seen ads 52.7 57.7 55.8 59.6 76.4 76.6 64.6 for cigarettes on billboards (3.5) (9.2) (3.5) (5.7) (4.6) (5.3) (5.1) Reported having seen ads for 48.7 72.6 64.0 51.7 80.7 74.7 66.7 cigarettes in newspapers or magazines (5.3) (6.8) (3.4) (3.6) (3.9) (4.9) (4.0) Reported having seen ads for 53.4 57.1 55.1 56.7 78.3 73.1 62.2 cigarettes at sporting and other events (6.3) (4.1) (3.8) (5.2) (5.3) (5.8) (6.2) Reported having been offered free 11.0 13.3 14.4 13.7 15.2 8.7 14.5 cigarettes by tobacco company (2.1) (2.5) (2.5) (2.9) (4.4) (3.7) (3.4) Reported having seen anti-smoking 69.0 83.6 85.7 65.9 79.8 80.7 69.7 media messages (4.8) (2.7) (2.5) (4.2) (2.8) (3.6) (6.1) Reported having been taught the 57.7 68.7 68.9 42.1 38.7 34.1 51.6 dangers of smoking in class (7.0) (5.2) (2.5) (4.8) (4.8) (5.9) (5.7) Reported having discussed reasons why 32.1 44.5 50.7 28.7 29.4 26.7 34.9 people their age smoke in class at school (5.0) (8.3) (4.1) (3.0) (4.3) (5.7) (5.5) Source: Global Youth Tobacco Survey Collaborative Group 2002. Note: Numbers in parentheses confidence interval, percent; n.a. not applicable (fewer than 35 participants answered this question); ETS environmental tobacco smoke. a. Currently smoking cigarettes or used other tobacco products on at least 1 of the previous 30 days. children identified the new flag. This showed clearly that a few had data on rates based on adequate national surveys this new South African symbol had been internalized by the (Corrao et al. 2000). These data are shown in table 18.2. age of five years. Therefore, it was of great concern when it Generally, men smoke much more than women do. With was found that between 51 and 76 percent of these children the exception of women in Guinea, Kenya, and Namibia, could identify the logos of popular cigarette brands. The only 20 percent or fewer of the women smoke. However, highest level of recognition was for the cigarette brand of these figures probably underestimate tobacco use, as ciga- the company that sponsored the soccer tournament in the rettes are not the only form of tobacco used in Sub-Saharan previous year. Seven percent of these five-year-old children Africa. A substantial number of people use hand-rolled cig- had experimented with cigarettes at least once (Thea de Wet, arettes or smoke a pipe, and the use of smokeless tobacco is unpublished data). common in women. In South Africa, Steyn and colleagues Only 23 of the Sub-Saharan Africa countries had any data (2002) found that 12.6 percent of black women used snuff, on cigarette-use prevalence rates available in 2000, and only a higher rate than the 5.3 percent of black women who Lifestyle and Related Risk Factors for Chronic Diseases | 249 Table 18.2 Prevalence of Adults Who Smoke Cigarettes, pro-smoking in orientation and on the verge of smoking, by Country whereas another 53 percent were found to have only a shal- Age in Year of Total Men Women low commitment to not smoking. Country years survey (%) (%) (%) It has been reported from South Africa that black women Benin 10 1988 37 -- -- start to smoke at a later age than do women elsewhere in the Botswana 15 1998 21 -- -- world (Steyn et al. 2002). The implication is that interven- Cameroon Adult 1994 37.5 -- -- tion programs in Sub-Saharan Africa should be targeting Chad 15 1994 -- 24.1 -- women in their reproductive years to prevent the onset of Congo, Dem. Rep. of 15 1998 -- 23.6 5.5 smoking and for smoking cessation programs. Côte d'Ivoire 15 1977 -- 42.3 1.8 Commercially manufactured cigarettes in Sub-Saharan Ethiopia Adult 1995 15.8 -- -- Africa are characteristically sold singly. It has been estimated Ghana 16 1980 -- 28.4 3.5 that in Nigeria up to 80 percent of cigarettes are sold in sin- Guinea 11­72 1998 -- 59.5 43.8 gles. Either people cannot afford to buy packs of 20 or they Kenya 20 1995 -- 66.8 31.9 sometimes supplement other forms of tobacco use with Lesotho 15 1992 -- 38.5 1 manufactured cigarettes (Saloojee 2000). Malawi Adult -- -- 20 9 Saloojee (2000) suggests that Sub-Saharan Africa has the Namibia Adult 1994 -- 65 35 lowest tobacco consumption rates and tobacco-related Nigeria 15 1998 -- 15.4 1.7 mortality in the world but that the current trend of tobacco Senegal 12­100 1998 -- 32 4.6 consumption implies that it is only a matter of time before Sierra Leone 15 1998 18.5 -- -- it succumbs to the same pattern of tobacco-related diseases South Africa 15 1998 -- 42 11 as the rest of the world. Projections suggest that the tobacco Swaziland Adult 1994 -- 24.7 2.1 epidemic can peak in Sub-Saharan Africa in the middle of Tanzania Adult -- -- 49.5 12.4 this century. Saloojee emphasizes that it is extraordinary to Uganda Adult 1995 -- 52 17 be able to predict an epidemic so far in the future and have Zambia Adult 1996 -- 35 10 the knowledge to prevent it. Price increases are the single Zimbabwe 15 1993 -- 34.4 1.2 most important factor in reducing smoking rates in a coun- try. The price elasticity, defined as the degree to which Source: Corrao et al. 2000. Note: -- data not available. tobacco consumption decreases with increasing price, is higher in developing countries than in richer countries. Significant price increases for cigarettes and other tobacco smoke regularly. Black women in Cape Town believe that products, therefore, would be a major step toward improved snuff has medicinal value as well as that of pain relief tobacco control for Sub-Saharan Africa (Saloojee 1995). (Marks, Steyn, and Ratheb 2001). Although smuggling may account for up to 50 percent of It is traditionally taboo for black women of Sub-Saharan consumption in some settings and may limit the impact of Africa to smoke during their reproductive years. Marks, price increases, economic policies have been effective in lim- Steyn, and Ratheb (2001) found that Xhosa women in Cape iting tobacco consumption. Town manifested this attitude, as 72 percent of the women participating in the survey thought it disgraceful, shameful, and taboo for them to smoke. The women who smoked did NUTRITION IN TRANSITION so in private settings or in secret. Almost all women who smoked said they did not smoke in front of their elders or A diet high in fat, particularly saturated fat, low in carbohy- parents. drates, fruit, and vegetables, along with a high salt intake In their promotion of tobacco products in Sub-Saharan leads to the emergence of chronic risk factors. Traditional Africa, tobacco companies target women and youth with diets in Sub-Saharan Africa, which are low in fat and high in images of "attractive, successful, upwardly mobile black unrefined carbohydrates, protect people against chronic women." They have thus overwhelmed many of the tradi- diseases. The nutrition transition refers to large shifts in tional taboos. This was revealed by Marks, Steyn, and the composition and structure of diets (Popkin 2001a). The Ratheb (2001), who found in their Cape Town study that dietary changes of the nutrition transition involve large 5 percent of nonsmoking black women were found to be increases in the consumption of fat (especially saturated fat) 250 | Krisela Steyn and Albertino Damasceno and sugar, marked increases in animal products, and a Figure 18.2 Macronutrient Intakes in South Africans, per decline in unrefined cereal and, thus, in fiber intakes Capita, 1993 and 1999 (Popkin 2001a, 2001b). 80 Nutrition patterns in Sub-Saharan Africa countries are 69.5 70 65.9 influenced by many factors, including individual preference; 60 culture, traditions, and beliefs; and price. However, avail- 50 ability and accessibility are the principal factors that shape 40 intake dietary patterns (N. P. Steyn, WHO, personal communica- % 30 tion). In the Sub-Saharan context, war and internal strife, 23.6 18.9 20 drought and poor agricultural practices, and rapid urbani- 11.6 11.2 10 zation are particularly influential. In addition, multinational 0 food companies market their products aggressively in the protein fat carbohydrate region. For example in West Africa, consumption has macronutrient changed from that of locally produced coarse grains, such as 1993 1999 millet and sorghum, to imported wheat and rice (Teklu Source: Steyn, Senekal et al. 2000; data extrapolated from food balance sheets. 1996). In the black population of Cape Town, it was found that a larger proportion of life spent in the city was associated with an increased consumption of fat and a decrease in car- Masai in Tanzania. Although their traditional diet was bohydrates (figure 18.1). This is reflected in an increased use high in cholesterol and fat from whole milk, blood, and of dairy produce, meat, fat, and nonbasic food items and a meat, and included little green vegetables or fruit, they decreased intake of cereals (Bourne, Lambert, and Steyn maintained low serum cholesterol levels (Gibney and 2002). Burstyn 1980). It was suggested that this unexpected rela- An examination of food balance sheets, based on the tionship could be explained by the cholesterol-lowering amount of food sold in South Africa indicates that per effect of the high levels of fermented milk consumed by the capita available fat consumption increased by 27 percent Masai and the substantial amounts of saponins and pheno- between 1993 and 1999 (figure 18.2). This increase was lic compound in the plant dietary additives used by them accompanied by a decrease in carbohydrate consumption. (Johns et al. 1999; St-Onge, Farnworth, and Jones 2000). An enigma in the proposed relationship of a diet high in Unfortunately, these protective dietary practices have cholesterol and fat, but with little fruit and vegetables, and declined and new dietary changes have been observed in the raised blood cholesterol levels was reported for the pastoral pastoral Masai; consequently there has been a significant increase in their serum total cholesterol levels (McCormick and Elmore-Meegan 1992). Not even the physically active Masai of Sub-Saharan Africa could escape the impact of the Figure 18.1 Fat and Carbohydrate Intake as Functions of nutrition transition. Proportion of Life Spent in a City 80 67.6 63.3 60.8 AEROBIC EXERCISE 60 57.8 Adequate physical activity has been shown to have many 40 energy health-promoting properties and has a direct, independent % 30 25 24.8 role in reducing cardiovascular disease mortality (Haskell, 22.5 20 14.4 Leon, and Caspersen 1992; McBride et al. 1992). Tradition- 13.8 13.8 15 ally, it has been thought that a high level of physical exercise 0 could in part explain the low levels of chronic diseases found 0 ­ 20 20­ 50 50­ 80 80­100 in Sub-Saharan Africa countries. However, the amounts of % life in city physical exercise have been decreasing as a result of the high % carbohydrate % fat % protein degree of urbanization that has been occurring across the Source: Bourne, Lambert, and Steyn 2002. continent. In urban settings, public transport replaces the Lifestyle and Related Risk Factors for Chronic Diseases | 251 traditional pattern of walking long distances, and urban Benefice 2001). The girls were spending a minimum of employment usually entails far less physical labor than rural 3.5 hours per day doing domestic duties, including collect- employment or other activities of daily living, such as chop- ing water and stomping maize. The level of physical activity ping wood, carrying water, or tilling the fields. In the cities, was higher during the rainy season. During the dry season, high crime levels prevent people from moving about freely. some of these adolescent girls were sent to the cities to work In the poorest periurban settings, inhabitants watch televi- as servants. It was found that the adolescents in the city had sion more frequently than their rural counterparts do. Few an even higher physical expenditure than those who stayed studies on the physical activity patterns of people in Sub- at home. However, despite the heavier workload, those in Saharan Africa have been published. the city had a better nutritional status. These healthy A study carried out in Cameroon comparing rural with Senegalese girls all had higher physical activity levels than urban people 15 years of age or older clearly illustrated those reported for girls in developed countries. lower rates of physical activity in the urban settings A health and fitness survey of adolescent schoolchildren (Sobngwi et al. 2002). Similar differences were found for age 12 to 18 years in the Western Cape in South Africa con- healthy, elderly people in Nigeria (Ezenwaka et al. 1997). In ducted by Lambert and colleagues in 2000 found that a high the urban settings, 62 percent of men and 83 percent of level of fitness was inversely associated with current BMI in women age 55 years or older led a sedentary lifestyle, where- both boys and girls. Inactivity, defined as watching televi- as this was the case for only 22 percent of men and 50 per- sion for more than three hours a day, was associated with cent of women in the rural areas. In Nigeria it was also current BMI and a low fitness level. The use of BMI projec- found that civil servants with high seniority had lower levels tion formulas to predict future BMI suggests that 24.5 per- of physical activity than their junior counterparts, suggest- cent of the females and 12.8 percent of the males will be ing that upward social mobility was associated with less overweight at the age of 18 years (Lambert et al. 2000). physical activity (Forrest et al. 2001). Two surveys carried out seven years apart in Maputo, Many studies across Sub-Saharan Africa have revealed Mozambique, of students age 8 to 15 years in the same group the impact of a sedentary lifestyle on emerging NCD risk of schools, showed a significant increase of mean systolic factors. The physical activity of the Nigerian civil servants blood pressure, diastolic blood pressure, and total body fat studied by Forrest and colleagues (2001) was mostly attrib- between the 1992 and the 1999 samples for all ages and for uted to occupational activities. Low levels of physical activ- both sexes but no changes in the mean level of cholesterol. ity were correlated to weight, body mass index (BMI), waist- These changes of cardiovascular risk factors were attributed to-hip ratio, blood pressure, insulin levels, and total and to the reductions in physical activity observed concomitantly low-density lipoprotein (LDL) cholesterol in men. Similarly, with changes in nutrition (Damasceno and Prista 2001). Sobngwi and colleagues (2002) in Cameroon as well as One of the reasons for the small number of published Aspray and colleagues (2000) in Tanzania concluded that studies on physical activity in Sub-Saharan Africa is the dif- physical inactivity was associated with obesity, diabetes, and ficulty of measuring it in large epidemiological studies. For hypertension in the people they studied in urban and rural such studies researchers have to rely on physical activity settings in both Cameroon and Tanzania. A qualitative study questionnaires that must be accurate, valid, and repro- in Cameroon also found that the reduced physical activity ducible. Some efforts have been made to develop question- accompanying sedentary occupations in the cities explained naires that may be useful for Sub-Saharan Africa countries the higher rate of obesity observed in people with these or that can be shown to be reliable in all regions of the sedentary occupations (Treloar et al. 1999). Levitt, Steyn, world. Sobngwi and colleagues (2001) clearly showed how Lambert, and colleagues (1999) showed an independent necessary this validation process is for the use of physical association between low levels of physical activity and hav- activity questionnaires with people in Sub-Saharan Africa, ing type 2 diabetes in a poor, periurban community near as study participants' self-ranking of their physical activity Cape Town. did not match the tested questionnaire's quartiles of physi- In an interesting three-year prospective study in Senegal, cal activity (Heini et al. 1996; Sobngwi et al. 2001). researchers followed the unusually high physical activity Wareham (2001) emphasized that to develop internation- level of adolescent girls between the ages of 13 and 15 years; ally standardized questionnaires, such as the International physical activity was measured by accelerometers recording Physical Activity Questionnaire, in all countries efforts will minute-by-minute movements (Benefice and Cames 1999; have to be made to ensure that validation and reliability Benefice, Garnier, and Ndiaye 2001a, 2001b; Garnier and studies take place. 252 | Krisela Steyn and Albertino Damasceno OBESITY pregnancy and lactation. It is therefore unsurprising that studies have shown that black women in South Africa also The World Health Organization (WHO) defines obesity as a do not perceive being overweight or obese as a health risk condition in which excess body fat has accumulated to (Ndlovo and Roos 1999). such an extent that health may be adversely affected. The Urbanization, associated with changing dietary patterns degree of body weight is usually expressed as BMI; this is and less physical activity and a rise in socioeconomic status, the ratio of weight in kilograms to the square of height in is occurring across Sub-Saharan Africa countries. All these meters. The BMI is used to classify a person's body weight as factors lead to an increase in the prevalence of overweight underweight (BMI less than 18.5), normal weight (BMI and obese people in the region. This phenomena is illus- 18.5­24.9), overweight (BMI 25­29.9), or obese (BMI trated in a study by Kruger and colleagues (2002) of the greater than 30) (WHO 2000). nutrition and physical activity patterns of a large sample of In addition, it is customary to indicate the amount of people from the North West Province of South Africa ex- abdominal fat mass. This can vary considerably among indi- posed to all levels of urbanization. The researchers found a viduals who have the same BMI. Abdominal fat is reported significant association between household income and by measuring the waist circumference or the waist-to-hip measures of obesity. They also saw a positive correlation circumference ratio. The waist circumference is thought to between total energy intake, fat intake, and BMI. The physi- provide a better correlate with abdominal fat mass than the cal activity index correlated negatively with BMI and waist waist-to-hip ratio. High abdominal fat mass is frequently circumference. referred to as central obesity. This form of obesity has been Cooper and colleagues (1997) also illustrated these shown to have more morbidity than if the fat distribution is trends when they compared the mean BMI of urban and predominantly on the hips (WHO 2000). rural West Africans with that of Jamaicans and blacks in the Obesity greatly increases the risk for conditions such as United States. A gradual increase in mean BMI was evident type 2 diabetes, hypertension, dyslipidemia, gall bladder along the historical migration path of the West African dias- disease, sleep apnea, osteoarthritis, and lower back pain. It pora that was attributed to a cross-cultural gradient of has also been shown to be associated with coronary artery decreasing physical activity and Westernization of the diet. disease and some cancers, and to reduce life expectancy. Table 18.3 shows the prevalence of obesity from Sub- Central obesity has been shown to be associated with meta- Saharan Africa reported in recent literature. South Africa has bolic syndrome. The key features of this condition are raised the highest prevalence of obesity in black men and women blood pressure, raised insulin and triglyceride levels, reported in Africa. Its prevalence among urban black women reduced high-density lipoprotein (HDL)-cholesterol levels, of 36 percent in South Africa already exceeds that of black and insulin resistance. The condition is strongly atherogenic women in the United States (Cooper et al. 1997; Puoane and predisposes to an elevated risk of diabetes and cardio- et al. 2002). The prevalence of obesity in South African and vascular disease (Fontaine et al. 2003; Peeters et al. 2003; urban Tanzanian men was very similar, whereas the preva- Solomon and Manson 1997; WHO 2000). lence of obesity in South African women was almost double In many Sub-Saharan Africa countries, an increased level that of urban women in The Gambia and Tanzania. of body fat is associated with beauty, prosperity, health, and An analysis of the mortality patterns associated with obe- prestige, despite its negative impact on health. Thinness, in sity in American people suggests that it has less of an impact contrast, is perceived to be a sign of ill health or poverty and on mortality and years of life lost in African Americans than is something to be feared and avoided, particularly in recent in other Americans (Fontaine et al. 2003; Solomon and years, when it has been associated with AIDS (Treloar et al. Manson 1997). Although no similar data are available on the 1999). In disadvantaged communities in South Africa, food association between obesity and mortality in Sub-Saharan is highly valued because food security has not always been Africa countries, it would be unwise to assume that obesity ensured. Researchers found it to be socially unacceptable for is benign for the people of this region. Reference has been an individual to refuse to eat food that was offered to them made to the impact that obesity has on the emergence of (Mvo, Dick, and Steyn 1999). Brown and Konner (1987) hypertension, diabetes, and other chronic disease risk fac- also reported that the majority of the less developed regions tors in Sub-Saharan Africa. Furthermore, high rates of LBW had, or still have, ideals of feminine beauty that include babies are common in the region, and this poses a risk to plumpness, which is consistent with the hypothesis that fat these babies in later life to develop obesity and other chronic stores function as a cushion against food shortages during disease risk factors. Lifestyle and Related Risk Factors for Chronic Diseases | 253 Table 18.3 Anthropometric Indicators, by Country Height Weight BMI Waist Hip Waist- Obese Population n (cm) (kg) (kg/m2) (cm) (cm) hip ratio (%) Body fat (%) Men Nigeria 1,171 168.3 61.5 21.7 77.3 88.3 0.88 -- -- Cameroona Urban 612 172.3 74.5 25.1 83.3 96.8 0.86 -- -- Rural 745 170.1 68.1 23.5 80.4 90.7 0.89 -- -- South Africa (black) 5,401 169 65.2 22.9 81.8 94.2 0.9 7.5 -- Cameroonb Urban 461 -- -- 25 -- -- -- -- -- Rural 308 -- -- 21 -- -- -- -- -- Zimbabwe (urban) 384 171.0 63.0 21.7 79.0 91.0 0.86 -- 21.1 Gambia, The Urban 1,028 -- -- 20.8 -- -- -- 1.8 -- Rural 1,200 -- -- 19.8 -- -- -- 0.1 -- Tanzania (urban) 3,659 -- -- -- -- -- -- 6.9 -- Women Nigeria 1,338 158.3 56.6 22.6 73.9 93.5 0.79 -- -- Cameroona Urban 749 162.1 71.0 27.0 82.5 102.5 0.81 -- -- Rural 722 160.7 60.6 23.5 80.9 92.6 0.87 -- -- South Africa (black) 7,726 158 67.8 27.1 85.5 104.6 0.8 30.0 -- Cameroonb Urban 591 -- -- 26.0 -- -- -- -- -- Rural 438 -- -- 22.0 -- -- -- -- -- Zimbabwe (urban) 391 160 67.0 21.7 82.0 101.0 0.82 -- 36.5 Gambia, The Urban 1,138 -- -- 23.9 -- -- -- 12.2 -- Rural 2,023 -- -- 20.5 -- -- -- 1.1 -- Tanzania (urban) 5,654 -- -- -- -- -- -- 17.4 -- Source: Cooper et al. 1997; Puoane et al. 2002; Mbanya et al. 1998; Mufunda et al. 2000; van der Sande et al. 2000; Bovet et al. 2002. Note: -- data not published; age-standardized against world population. a. Data from Cooper et al. 1997. b. Data from Mbanya et al. 1998. HYPERTENSION Black hypertensive patients in Sub-Saharan Africa are prone to cerebral hemorrhage, malignant hypertension, High blood pressure is a major risk factor for heart attacks kidney disease leading to uremia and congestive heart and strokes. It also contributes to renal disease and blind- failure, whereas coronary heart disease is relatively ness. It is estimated that between 10 million and 20 million uncommon. Responses to antihypertensive medication people in Sub-Saharan Africa have hypertension. It has been drugs like the beta-blockers and the angiotension- estimated that adequate hypertension treatment of these converting enzymes (ACE) inhibitors are poor unless people could prevent about 250,000 deaths (Cappuccio these agents are combined with a thiazide diuretic. Black et al. 2000). However, hypertension in Sub-Saharan Africa is patients respond best to diuretics, vasodilators or calcium universally under diagnosed or inadequately treated, or channel blockers. both, with the result that extensive end-organ damage and premature death are often seen. Furthermore, hypertension Kaufman and colleagues (1996) reported that the risk of frequently co-exists with other NCD risk factors, such as death increased by 60 percent with an increase of 20 mmHg diabetes. Seedat (1999) summarized the pathophysiology of in diastolic blood pressure in rural Nigeria. They estimated hypertension and response to treatment as follows: that the population-attributable risk or the reduction in 254 | Krisela Steyn and Albertino Damasceno mortality that would have been observed if hypertension South Africa, that the duration of urbanization independ- were not present was 7 percent, showing the impact of ently predicted the presence of hypertension. hypertension on all-cause mortality in rural Nigeria. Table 18.4 shows the prevalence rate in some Sub- Malignant hypertension also occurs more frequently in Saharan Africa countries from studies published since 1997. black people of Sub-Saharan Africa than in other ethnic The studies reveal that although some countries still main- groups. Milne and colleagues (1989) showed this in black tain large differences between urban and rural prevalence, patients hospitalized for hypertension. This condition is fre- differences are no longer apparent in many countries. The quently related to severe renal disease and hypertensive prevalence rates in the rural areas have increased to levels encephalopathy. Data from the South African Dialyses and similar to those found only in the cities in the past. As an Transplantation Registry have shown that hypertension was example, in the early 1990s Mollentze and colleagues (1995) responsible for 35 percent of end-stage renal failure in showed that the rural community of QwaQwa in the Free blacks and that malignant hypertension was diagnosed in State, South Africa, had rates of hypertension similar to 57 percent of the black patients with essential hypertension those in the periurban community of Mangaung in the (Veriava et al. 1990). Untreated malignant hypertension has same province. The previously observed differences in been shown to have a five-year survival as low as 1 percent. hypertension prevalence between poorer and more affluent Earlier surveys showed that the lowest prevalence of countries have also diminished. hypertension occurred in the poorest Sub-Saharan Africa Some dietary factors are related to hypertension, includ- countries, and as affluence increased, the prevalence ing increased salt (sodium) intake and a decrease in fruit and increased. The surveys also revealed that hypertension was vegetables (potassium); a higher intake of alcohol products, more common in urban than in rural settings in the region particularly by men, also plays a role. The association (Nissinen et al. 1988). The elegant Kenyan Luo migration between hypertension and obesity has been well documented study of Poulter and colleagues (1990) was the first to show in many countries in Sub-Saharan Africa. In Zimbabwe, that migration of people living in traditional rural villages Mufunda and colleagues (2000) found this strong associa- on the northern shores of Lake Victoria to the urban settings tion, as did Rotimi and colleagues (1995) in populations of of Nairobi was associated with an increase in blood pres- West African descent. Despite this clear association it has sure. This suggests a marked change in the diet of the new been suggested that the noxious effect of obesity in black arrivals in Nairobi to a higher salt and calorie intake along people is less than in people of other ethnic groups. Most of with a reduced potassium intake due to consuming less fruit the supporting evidence for this viewpoint is based on stud- and vegetables. Studies carried out in Nigeria that compared ies carried out with African Americans in the United States. urban and rural people's sodium and potassium excretions A small study in South Africa suggested similar findings observed similar findings (Kaufman et al. 1999). Higher (Walker et al. 1990). The influence of alcohol consumption, pulse rates in the Nairobi participants suggest that mecha- particularly heavy drinking, on increasing blood pressure nisms related to increased autonomic nervous system levels has also been described in Nigeria (Bunker et al. 1992; activity could contribute to the higher levels of blood pres- Ekpo et al. 1992). The data on the association between high sure observed (Poulter et al. 1985, 1990). Urban Nigerians salt (sodium chloride) intake and hypertension in black reported higher stress levels and lower social integration people from Africa has been summarized by Seedat (1996) scores than their rural counterparts. These indicators of and suggests that black people have a transport mechanism increased stress in the urban setting were associated with of high sodium retention and a low rennin activity. Mtabaji higher blood pressures (Kaufman et al. 1999). and colleagues (1992) found salt sensitivity, measured by the Socioeconomic status and urbanization are good predic- blood pressure response on salt loading, in 46.2 percent of tors of hypertension. Zimbabwean women doing traditional study subjects in Tanzania. A high intake of sodium is com- work-related activities on rural communal land had lower mon in Sub-Saharan Africa, as it is used to preserve food blood pressures than did those women who were working or to make food tastier. For example, Cappuccio and col- for a wage on large-scale, commercial agricultural farms. leagues (2000) described the diet in Ghana as consisting The latter group, in turn, had lower blood pressures than mostly of unprocessed food and highly salted fish and meat. women who earned a living in more industrial mining areas Substantial amounts of salt are added to food while cooking, (Hunter et al. 2000). Similarly, researchers (Steyn, Fourie and monosodium glutamate­based flavoring cubes or salts et al. 1996) found in the black community of Cape Town, are widely used to give food taste. In addition to a high salt Lifestyle and Related Risk Factors for Chronic Diseases | 255 Table 18.4 Large Prevalence Studies on Hypertension since 1997, by Country Cut-off point Cut-off point 140/90 mmHg 160/95 mmHg Males Females Males Females Equipment for blood Country Age (years) (n) (%) (n) (%) (%) (%) pressure measurement Nigeria Urban (adults; -- 581a -- 417a -- 13.9 5.0 Mercury sphygmomanometer civil servants) Urban 25­55 1,171 14.7 1,338 14.3 6.9 6.9 Mercury sphygmomanometer Cameroonb Urban 25­55 612 22.8 749 16.0 8.7 8.7 Mercury Rural 25­55 745 14.2 722 16.3 4.7 7.4 sphygmomanometer Cameroon Urban 25­74 461 -- 591 -- 16.4 12.1 Mercury Rural 25­74 308 -- 438 -- 5.4 5.9 sphygmomanometer Tanzania Urban (Ihala) 15 330 30.0 437 28.6 14.4 15.4 Mercury Rural (Shari) 15 401 32.2 527 31.5 17.2 15.6 sphygmomanometer Tanzaniab Urban 35­64 1,729 27.1 2,053 30.2 13.1 17.7 Visomat Gambia, The Urban (Banjul) Adults 1,028 22.0 1,138 16.9 7.5 7.3 Omron HEM-705-CP Rural (Farafenni) Adults 1,200 20.6 2,023 16.0 7.6 6.3 Zimbabwe 25 384 28.0 391 41.0 18.0 28.0 Omron HEM-713C Urban South Africab African 15­65 4,283 23.5 6,174 25.0 12.9 15.7 Omron M1 Coloured 15­65 772 27.3 1,008 29.6 14.9 22.7 electronic blood White 15­65 500 33.5 603 20.6 23.3 18.5 pressure manometer Asian 15­65 183 28.0 279 25.3 18.7 20.5 Source: Olatunbosun et al. 2000; Cooper et al. 1997; Mbanya et al. 1998; Edwards et al. 2000; Bovet et al. 2002; van der Sande et al. 2000; Mufunda et al. 2000; Steyn et al. 2001. Note: -- data not published; adjusted to the world population. a. No rates. b. Adjusted to the world population. intake, people in Sub-Saharan Africa frequently eat little These data clearly suggest that many nutritional inter- fruit and vegetables, resulting in low potassium intakes. ventions are required to decrease the prevalence of hyper- Few intervention studies have been conducted in Sub- tension. These interventions include regulation of the Saharan Africa to show that a reduction in salt and an amount of salt used by the food industry in Sub-Saharan increase in potassium improve the blood pressure in its Africa, the promotion of increased potassium consumption populations. A study done in Tanzania showed that a low- in the form of fruit and vegetables, and the limitation of sodium diet leading to a low urinary excretion level of alcohol use. 52 millimoles per day reduced blood pressure in normoten- Despite the many environmental factors related to hyper- sive people significantly within four to five days (Mtabaji, tension, many studies in Sub-Saharan Africa suggest a Nara, and Yamori 1990). A study in Kenya reported that sup- possible genetic contribution to the origins of hypertension plementation with potassium in newly diagnosed patients in black people. Van der Sande and colleagues (2001) in The with hypertension reduced the blood pressure to a level Gambia and Steyn and colleagues (unpublished data) in similar to that found in patients treated with a diuretic South Africa have found that high blood pressure is associ- (Obel and Koech 1991). ated with a strong family history of either hypertension or 256 | Krisela Steyn and Albertino Damasceno stroke. This could provide a cost-effective opportunity to Boerwinkle 2000; Nkeh et al. 2003), or have been shown to identify people who need more detailed screening for hyper- occur with too low a frequency to contribute substantially to tension. Rotimi and colleagues (1999) used computer mod- population-attributable risk (Barlassina et al. 2000). Family- els and regression analyses to estimate the degree of heri- based linkage studies assessing blood pressure as a continu- tability of systolic and diastolic blood pressure in Nigerian ous variable and using 24-hour ambulatory blood pressure families. The heritability estimate was 45 percent and 43 per- monitoring techniques are under way in South Africa to fur- cent for systolic and diastolic blood pressure, respectively. ther evaluate the role for these variants as determinants of This emphasizes interaction between environmental influ- blood pressure in groups of African ancestry. ences and genetic factors in the etiology of hypertension. Effective management of hypertension usually requires In populations of African ancestry, data from small case- treatments with more than one drug. In Ghana, Hesse and control studies and a large case-control study conducted in Nuama (1997) found that only 18 percent of a group of South Africa with more than 700 cases phenotyped using patients with hypertension had one drug prescribed, where- 24-hour ambulatory blood pressure monitoring and 700 as 60 percent had two drugs and 22 percent had three or controls (G. R. Norton et al., personal communications) more drugs prescribed. The use of two or more drugs will suggest that the angiotensin-converting enzyme (ACE) gene inevitably result in a high cost for antihypertensive medica- variant contributes little to hypertension. Nevertheless, tion, especially when newer medications are used. However, gender-specific effects need to be excluded. In addition, there are cheaper, older, and effective medications available angiotensinogen (AGT) gene variants that have been associ- in resource-scarce settings in Sub-Saharan Africa. Hesse and ated with hypertension in Caucasian groups are not impli- Nuama (1997) also reported that between 1973 and 1993 a cated in subjects of African ancestry (Tiago et al. 2002). diuretic was the type of drug prescribed for initial treatment However, an alternative functional promoter region variant of patients with hypertension; diuretics were used in 90 per- of the AGT gene ( 217GA) has recently been shown to be cent of cases, including reserpine in 46 percent, methyldopa strongly associated with hypertension in a small African in 31 percent, and propranolol in 30 percent. With the American case-control study (Jain et al. 2002), data that has exception of methyldopa all these medications are inexpen- now been confirmed in a large case-control study conduct- sive and suited to resource-scarce settings. ed in black South Africans (G. R. Norton et al., personal The benefits of treating hypertension have been convinc- communication). The role of the 217GA AGT gene ingly shown in many parts of the world, but relatively little variant in contributing to the variance of blood pressure data are available from Sub-Saharan Africa. One example is within families is now being assessed in a South African a study by Salako and colleagues (1999), who showed that study. Furthermore, an additional functional AGT gene pro- when patients with hypertension are given effective treat- moter region variant ( 20AC; Zhao et al. 1999) has been ment, there is a reduction in the excretion of urinary albu- shown to modify the impact of body size on blood pressure min, suggesting improved renal function in patients who (Tiago et al. 2002), and the 217GA variant's effect on previously had early renal impairment. the risk for hypertension (G. R. Norton et al., personal After reviewing the available data, Cooper and colleagues communications) in subjects of African ancestry, thus (1998) suggested: "Hypertension is fully treatable, but social suggesting complex genotype-genotype and genotype- conditions in Africa make the implementation of blood phenotype interactions of the AGT gene in people of pressure control programs difficult. Lack of a clear strategy African origins. based on evidence has undermined these efforts." They also Other candidate gene variants implicated in blood estimated that effective hypertension treatment would pressure control or hypertension in those of European or lead to a reduction in population-attributable risk of 2 per- African ancestry, including functional or potentially func- cent in Africa compared with 0.15 percent in the United tional variants found within the guanosine triphosphate States. "Number needed to treat" analyses showed that protein 3 subunit gene (Siffert et al. 1998), the sodium the cost of drugs to prevent one death is US$1,800 in Africa, epithelial channel gene (Baker et al. 1998), the -adducin if the cheaper drugs are used, whereas it is US$14,000 to gene (Cusi et al. 1997), and the 2 receptor gene (Svetkey US$1 million in the United States, depending on which et al. 1996), have been shown not to be associated with drugs are used. Cooper and colleagues (1998) concluded hypertension in all studies conducted in groups of African that the treatment of hypertension should be a health prior- ancestry (Candy et al. 2000; Larson, Hutchinson, and ity in Sub-Saharan Africa. Lifestyle and Related Risk Factors for Chronic Diseases | 257 DYSLIPIDEMIA Johannesburg and Soweto, Steyn and colleagues (2000) found that the mean total cholesterol level in African and Dyslipidemia is defined as a clinically significant alteration coloured five-year-old children was 3.9 millimoles per liter of the naturally occurring blood lipids and lipoproteins compared with 4.1 millimoles per liter for Indian and predisposing to cardiovascular diseases and other chronic 4.4 millimoles per liter for white children. The correspon- diseases (Berger and Marais 2000). The two commonly ding ratios of HDL cholesterol to total cholesterol were measured blood lipids are cholesterol and triglycerides. 30.6 percent, 28.9 percent, 27 percent, and 25.8 percent, These molecules are carried in a range of lipoprotein parti- respectively. These findings in children at the age of five cles in the blood. The most important lipoprotein particles years correspond to the blood lipid pattern found in people that predispose to atherosclerosis and thus cardiovascular of African descent in most studies carried out across Africa. diseases are the chylomicrons high in triglyceride and These group differences in lipid profiles tend to continue LDL particles, which carry mainly cholesterol. The LDL par- into adolescence and adulthood. The lower total cholesterol ticle is the most atherogenic of all the lipoprotein particles. and higher HDL cholesterol levels of persons of African The HDL particle protects the arteries against atherosclero- descent indicate a protective pattern against developing sis and, consequently, against cardiovascular diseases. HDL atherosclerosis and, consequently, against ischemic heart transports cholesterol away from the arteries to be excreted disease (Steyn et al. 2000). Seftel and colleagues (1993) com- via the liver. The most recent international recommenda- pared the lipid levels of male students between the ages of 15 tions regarding lipid levels are that a total blood cholesterol and 20 years from different ethnic groups in South African level below 5.2 millimoles per liter is desirable; similarly, a urban and rural settings. The mean total cholesterol level level of LDL cholesterol below 2.6 millimoles per liter and was 3.2 millimoles per liter for the rural and 3.7 millimoles a triglyceride level below 1.7 millimoles per liter are per liter for the urban students, differences that were statis- considered optimal. Persons with blood lipids below these tically significant. The rural black males age 15 to 20 had the levels do not carry a risk of developing atherosclerosis. An lowest total cholesterol level of all the groups studied. Only HDL cholesterol level below 1 millimole per liter also pre- 1 percent of the rural and 2 percent of the urban black males disposes to atherosclerosis development, whereas a level had total cholesterol levels above 5.2 millimoles per liter. above 1.6 millimoles per liter provides protection. The ratio In 1980 Knuiman, Hermus, and Hautvast (1980) reported of HDL cholesterol to the total cholesterol level, expressed as even lower total cholesterol levels in rural and urban boys a percentage, is frequently reported. A level of more than age seven to eight years in Côte d'Ivoire, Ghana, and Nigeria. 20 percent for men and 25 percent for women is considered The mean total cholesterol levels in these boys ranged from to provide protection against developing atherosclerosis 2.6 millimoles per liter in rural Nigeria to 3.5 millimoles per (Adult Treatment Panel III 2001). liter in urban Ghana, and their ratio of HDL cholesterol to A particularly atherosclerotic combination of lipid and total cholesterol was high, varying from 30 to 36 percent. other risk factors is referred to as the metabolic syndrome Walker and Walker (1978) also reported high levels of HDL and leads to premature coronary heart disease. This cholesterol in African children and adults in a population syndrome consists of high levels of blood triglyceride, low free of coronary heart diseases. levels of HDL cholesterol, and small dense LDL particles. It is Seftel, Raal, and Joffe (1995) reviewed the early studies on also associated with type 2 diabetes, hypertension, abdomi- total cholesterol levels in South African adult populations. nal obesity, insulin resistance, and physical inactivity.Van der They found that the mean total cholesterol in middle-aged Sande and colleagues (2000) found significant numbers of black men was about 4 millimoles per liter compared with urban black people in Banju, the capital of The Gambia, with 5 millimoles per liter in coloured men and 6 millimoles per the coexistence of these conditions, suggesting that the meta- liter in white men. Since the mean total cholesterol concen- bolic syndrome is common in this city. trations of newborn black and white babies were almost the Since the level of ischemic heart disease reported in same, the differences in middle-aged men were likely to be people from Africa is low, few hypertension and diabetes environmental in origin (Seftel, Raal, and Joffe 1995). surveys in the region include blood lipid measurements. The The findings of some recent lipid studies in Sub-Saharan available studies suggest that most people of African descent African countries are shown in table 18.5. A study of 8,581 have far lower blood lipid levels than found in people of rural Tanzanians, age 15 years and older, found significantly European or Indian descent in the region. These differences higher mean total serum cholesterol levels in women than among the groups are present from a young age. In in men (Swai et al. 1993). Only in the economically more 258 | Krisela Steyn and Albertino Damasceno Table 18.5 Mean Lipid Levels and Prevalence of Dyslipidemia in Black Subjects Total cholesterol Triglycerides Total cholesterol Triglycerides Location (age in years) n mmol/L HDL-C mmol/L mmol/L 5.2 mmol/L (%) 1.7 mmol/L (%) Tanzania Rural males 93 3.4 1.0 -- 4.4 -- Rural females 91 4.4 1.0 -- 22.5 -- Nomadic males 41 5.1 1.1 -- 48.6 -- Nomadic females 61 5.3 1.3 -- 53.7 -- Urban males 81 4.7 1.2 -- 29.5 -- Urban females 79 5.3 1.3 -- 50.0 -- Kilimanjaro (rural) Males (55 ) 413 -- -- -- 26.9 12.6 Females (55 ) 478 -- -- -- 35.5 11.5 Mara Males (55 ) 198 -- -- -- 6.6 8.1 Females (55 ) 133 -- -- -- 6.1 12.2 Morogoro Males (55 ) 79 -- -- -- 12.7 6.3 Females (55 ) 74 -- -- -- 9.5 12.1 Nigeria Males 110 -- -- -- 7.3 -- Females 36 -- -- -- 19.5 -- Gambia, The (normotensive) Urban males 1,028 4.1 -- 0.68 12.5 4.0 Urban females 1,138 4.6 -- 0.68 29.1 0.6 Rural males 1,200 3.6 -- 0.81 2.3 4.3 Rural females 2,023 3.9 -- 0.76 8.3 2.1 Gambia, The (hypertensive) Urban males 77 4.9 -- 1.13 24.5 4.1 Urban females 85 5.0 -- 0.79 39.1 4.7 Rural males 91 4.0 -- 1.17 4.1 1.4 Rural females 127 4.4 -- 0.83 13.7 2.9 South Africa Cape Town (periurban) 986 4.11 1.3 1.2 19.4 -- Free State QwaQwa (rural) 853 4.8 1.2 1.5 32.7 -- Mangaung (periurban) 758 5.3 1.4 1.2 37.4 -- Sources: Kadiri and Salako 1997; Mollentze et al. 1995; Njelekela et al. 2001; Oelofse et al. 1996; Swai et al. 1993; and Van der Sande et al. 2000. Note: -- data not published; HDL-C high-density lipoprotein cholesterol; mmol/L millimole/liter. advanced region, Kilimanjaro, did the total cholesterol levels that during the decade, a significant rise occurred in the increase with age as found in all Westernized people. The mean levels and in the prevalence of many cardiovascular favorable lipid profile in most of Africa may be threatened by disease risk factors, including increases in total cholesterol the march of time. Njelekela and colleagues (2001) surveyed levels and the prevalence of hypercholesterolemia. people for chronic disease risk factors in three settings in Two other cross-sectional studies in Tanzania assessed Tanzania in 1987 and again in 1998. Their findings suggest the impact of nutritional factors on the lipid profiles of Lifestyle and Related Risk Factors for Chronic Diseases | 259 people living in different settings. One compared the impact Westernized lifestyle and who have more than one chronic of the low-salt fish and vegetable diet on the lipid profile of disease risk factor. Although limited data are available on the Tanzanians with that of the typical diets consumed by peo- treatment status of high blood cholesterol in Sub-Saharan ple in Brazil and Italy (Pavan et al. 1997). The Tanzanians' Africa, a survey of about 13,000 patients of general practi- total blood cholesterol and their BMI were lower than that tioners in private practice in South Africa revealed that high of the other two groups. Another study compared the lipid blood cholesterol is poorly treated in those who could ben- profiles of Tanzanian villagers who consumed a diet consist- efit the most. These were 18.7 percent of men and 10.4 per- ing predominantly of fish with those of villagers who were cent of women who had ischemic heart disease in the past predominantly vegetarian (Pauletto et al. 1996). The total and whose total blood cholesterol levels were found to be blood cholesterol levels were lower (3.5 millimoles per liter) 5.9 millimoles per liter and 6.0 millimoles per liter, respec- in the fish-eating community than in the vegetarian com- tively. These patients need to lower their total blood choles- munity (4.1 millimoles per liter). Similarly, the blood terol levels significantly for protection against a recurrence triglyceride levels were 0.9 millimoles per liter and 1.3 mil- of heart disease. This is an achievable objective with the use limoles per liter, and the lipoprotein little (a) levels were 201 of highly effective newer cholesterol-lowering agents (Steyn milligrams per liter and 321 milligrams per liter in the fish- et al. 1998). eating community and vegetarian community, respectively (Pauletto et al. 1996). This study suggests that fish-eating communities have a more protective lipid profile than vege- HEALTH SERVICES' REQUIREMENTS FOR THE tarian ones. MANAGEMENT OF NCD RISK FACTORS Total cholesterol levels can be favorably influenced by improved dietary changes; however, increased physical Chronic disease rates are already higher than expected in activity patterns also improved the lipid profile. Researchers Sub-Saharan Africa countries. Consequently, these patients in Nigeria found that patients with hypertension, exercising are making significant demands on the health services. three times a week for 30 minutes per session over a 16-week Appropriate planning to manage chronic diseases and their period, decreased their total cholesterol and LDL- risk factors is of paramount importance in the region cholesterol levels slightly and increased their HDL- (Unwin et al. 2001). Most of the risk factors emerge during cholesterol levels (Iyawe, Ighoroje, and Iyawe 1996). middle age as a result of an unhealthy lifestyle that has been Although high total blood cholesterol levels are uncom- followed for several decades. Many patients have several mon in Sub-Saharan Africa countries, there are some excep- risk factors. In Nigeria it was found that patients may have tions. One example is provided by the descendents of the as many as five chronic disease risk factors, as shown in colonists from the Netherlands who came to the southern table 18.6 (Ezenwaka et al. 1997). The risk factors have a tip of Africa in the seventeenth century and belong to the synergistic effect on the total chronic disease risk. Health Afrikaans-speaking white community in South Africa. The services planned for prevention and care must therefore take phenomenon referred to by geneticists as the "founder effect" is reflected in the high concentration of familial hypercholesterolemia in this group. A small group of farm- Table 18.6 The Clustering of NCD Risk Factors in Nigerian Subjects ers from the Cape province trekked north into the interior (percent) regions of what later became South Africa, and tended to intermarry. The background prevalence of heterozygous Number of Total participants Men Women risk factors n 504 n 295 n 209 familial hypercholesterolemia across the globe is 1 to 500. However, the overall prevalence of familial hypercholes- 0 29.8 31.5 27.3 terolemia in rural, white, Afrikaans-speaking people in 1 70.2 68.5 72.7 South Africa was estimated to be 1 to 72 (Steyn, Goldberg 2 47.6 44.4 52.2 et al. 1996). 3 27.8 25.1 31.7 The treatment of high total blood cholesterol levels with 4 16.2 12.2 22.0 medication is most cost-effective in people who have the 5 9.2 6.4 13.2 highest risk of ischemic heart disease. These are people who Source: Ezenwaka et al. 1997. have had a previous heart attack or stroke or who suffer Note: Risk factors considered were smoking, alcohol intake, sedentary lifestyle, hypertension ( 160/95 mmHg), obesity, truncal obesity, hyperinsulinemia, insulin resistance, diabetes, from familial hypercholesterolemia and those with a typical hypercholesterolemia ( 5.2 mmol/L) and hypertriglyceridemia ( 2.5 mmol/L). 260 | Krisela Steyn and Albertino Damasceno cognizance of the burden of multiple chronic disease risk POLICY INITIATIVES TO PROMOTE factors in the same patient. A HEALTHY LIFESTYLE Data collected in large cohort studies conducted in the United States and the United Kingdom were used to assess The need for preventive action to reduce or prevent the the relationship between risk factors and cardiovascular dis- adoption of a less healthy lifestyle by people in Sub-Saharan ease (CVD) events. The studies used were the Framingham, Africa has never been greater. Many factors contribute to the Massachusetts, cohort data and United Kingdom cohort lack of chronic disease preventive programs being adopted data (ATP III 2001; Grundy et al. 1999; Joint Task Force, by governments (Yach 2002). 1998). Formulas were calculated that could be used to esti- The global health community has taken some significant mate a patient's chronic disease risk. These formulas usually steps to support countries in Sub-Saharan Africa and else- express the patient's risk of suffering from a CVD event in where to promote a healthy lifestyle for the entire world. The the next 10 years or longer period. Health services are thus WHO, a branch of the United Nations, spearheaded this ini- able to identify those patients who are at the highest risk of tiative by negotiations to propose a framework convention developing CVD and would benefit the most by using on Tobacco Control between 1999 and 2003. The member appropriate medication. Gaziano, Steyn, and Opie (2001) states of the WHO (2003) unanimously endorsed the frame- have shown that such an approach is more cost-effective work convention of Tobacco Control to the WHO's general than that of using single cut-off points for single CVD risk assembly in June 2004 for endorsement. The treaty attacks factors. Therefore, the total CVD risk-assessment approach tobacco use in many ways. It expects countries to provide clearly should be the one employed in settings with scarce treatment for people who smoke, to encourage cessation, and resources, as is the case in Sub-Saharan Africa countries. to prevent the onset of tobacco use in the young. It requires The model of health service provision prevailing across countries to protect the public from exposure to environ- most of Sub-Saharan Africa is one that focuses on acute ill- mental tobacco smoke and expects them to ban advertising to nesses, the immediate needs of patients, and episodic inter- the extent that their constitutions allow with respect to the action between the patient and the health services. This protection of commercial free speech. Thirty percent of any model does not provide adequate care for patients with tobacco product's packaging should contain health warnings chronic diseases. Consequently, millions of patients with of the danger of tobacco use. Misleading words such as"light" chronic diseases in the region are undiagnosed or receiving or "mild" should also be banned. It encourages countries to inadequate treatment. The WHO conducted a two-year raise taxes to the point that will discourage smoking and review of health care models and best practices from around to pass laws that hold tobacco companies accountable for the world. The findings, published in 2002, provided a com- medical and other costs incurred because of tobacco use prehensive conceptual framework for the prevention and (WHO 2003). With the exception of South Africa, very few management of chronic conditions in poorly resourced set- Sub-Saharan Africa countries have adequate tobacco control tings (WHO 2002). legislation in place to effectively protect their populations The WHO report focuses on the need to move away from against tobacco use. This treaty will encourage many coun- the acute model of care to efficiently coordinated and tries in the region to initiate such a development. patient-centered care. Such a move should facilitate an In March 2003 the WHO and the Food and Agriculture ongoing relationship between provider and patient and help Organization, another branch of the United Nations, issued patients to make full use of their own and their communi- a scientific report reviewing the data on diet, physical activ- ty's resources (Holman and Lorig 2000; Wagner 2000). This ity, and health. The WHO has embarked on an extensive approach suggests that the focus should be on the patients process of consultations across the globe, including Sub- in their own social settings and context and not only on the Saharan Africa countries, to ensure that a global strategy disease of the patient. This is of particular significance in provides realistic guidelines for all countries. The strategy Sub-Saharan Africa countries, where the patients' cultural will include a broad series of nutrition-related recommen- beliefs and practices should be understood when prevention dations, including dietary advice along with proposals deal- and management of chronic conditions are introduced. The ing with the labeling and advertising of food. Collaboration partnership between the patient and the health care with the food industry is seen as pivotal to improving the provider is not just a resource for understanding ill health, it quality of manufactured food and processed food. is the basis for the prevention and management of chronic All these initiatives will provide countries with informa- diseases and their risk factors (Swartz and Dick 2002). tion to identify and act appropriately to promote a healthy Lifestyle and Related Risk Factors for Chronic Diseases | 261 lifestyle. However, it is necessary to prioritize chronic disease ------. 1994. Mothers, Babies and Diseases in Later Life. London: British Medical Journal Publishing Group. prevention in view of all the other large health demands that Barlassina, C., G. R. Norton, N. J. Samani, A. J. Woodiwiss, G. C. Candy, face the ministries of health in Sub-Saharan Africa coun- I. Radevski, L. Citterio, G. Bianchi, and D. Cusi. 2000. "Alpha-Adducin tries. Countries with multiple burdens of disease will have to Polymorphism in Hypertensives of South African Ancestry." American develop priority-setting programs in order to plan how to Journal of Hypertension 13: 719­23. Benefice, E., and C. Cames. 1999. "Physical Activity Patterns of Rural use scarce resources most effectively. The failure of countries Senegalese Adolescent Girls during the Dry and Rainy Seasons to adopt the necessary steps to promote a healthy lifestyle, Measured by Movement Registration and Direct Observational however difficult such a decision might be early in the twen- Methods." European Journal of Clinical Nutrition 53 (8) : 636­43. ty-first century, will inevitably lead to increasing levels of Benefice, E., D. Garnier, and G. Ndiaye. 2001a. "Assessment of Physical Activity among Rural Senegalese Adolescent Girls: Influence of Age, obesity, hypertension, hyperlipidemia, and diabetes in the Sexual Maturation, and Body Composition." Journal of Adolescent populations of Sub-Saharan Africa countries. This, in turn, Health 28 (4): 319­27. will be the cause of an avoidable chronic disease epidemic ------. 2001b. "High Levels of Habitual Physical Activity in West African Adolescent Girls and Relationship to Maturation, Growth, and within a few decades. Nutritional Status: Results from a 3-Year Prospective Study." American Journal of Human Biology 13 (6): 808­20. Berger, G. M. B., and A. D. Marais. 2000. "Diagnosis, Management and CONCLUSION Prevention of the Common Dyslipidaemias in South Africa--Clinical Guideline, 2002." South African Medical Journal 90: 164­78. Bourne, L. T., E. V. Lambert, and K. Steyn. 2002. "Where Does the Black On reviewing the extent of chronic diseases in the low- Population of South Africa Stand on the Nutrition Transition?" Public income and middle-income countries, the Institute of Health and Nutrition 5 (1A): 157­62. Medicine of the United States National Academy of Sciences Bovet, P., A. G. Ross, J. P. Gervasoni, M. Mkamba, D. M. Mtasiwa, recommended that, in addition to the preventive actions C. Lengeler, D. Whiting, and F. Paccaud. 2002. "Distribution of Blood Pressure, Body Mass Index and Smoking Habits in the Urban required to reduce the risk factors, health services should Population of Dar es Salaam, Tanzania, and Associations with diagnose and control hypertension and diabetes and should Socioeconomic Status." International Journal of Epidemiology 31: 240­47. ensure access to low-cost drugs and the development of Brown, P. J., and M. Konner. 1987. "An Anthropological Perspective on affordable clinical care algorithms. The report also empha- Obesity." Annals of the New York Academy of Sciences 499: 29­46. sized the need to build capacity to conduct research and Bunker, C. H., F. A. Ukoli, M. U. Nwankwo, J. A. Omene, G. W. Currier, development activities and to develop institutional frame- L. Holifield-Kennedy, D. T. Freeman, E. N. Vergis, L. L. Yeh, and L. H. Kuller. 1992. "Factors Associated with Hypertension in works that would facilitate the development of chronic dis- Nigerian Civil Servants." Preventive Medicine 21 (6): 710­22. ease prevention activities and the transformation of the Candy, G., N. Samani, G. Norton, A. Woodiwiss, I. Radevski, A. Wheatley, health care services to adequately cope with the enormous, J. Cockcroft, and I. P. Hall. 2000. "Association Analysis of Beta2 Adrenoceptor Polymorphisms with Hypertension in a Black African mostly unacknowledged burden of chronic diseases in Sub- Population." Journal of Hypertension 18: 167­72. Saharan Africa (Howson et al. 1998). Cappuccio, F. P., J. Plange-Rhule, R. O. Phillips, and J. B. Eastwood. 2000. "Prevention of Hypertension and Stroke in Africa." Lancet 356: 677­78. REFERENCES Ceesay, S. M., A. M. Prentice, T. J. Cole, F. Foord, E. M. E. Poskitt, L. T. Weaver, and R. G. Whitehead. 1997. "Effects on Birth Weight and Perinatal Mortality of Maternal Dietary Supplements in Rural Gambia: Adult Treatment Panel III (Expert Panel on Detection, Evaluation, and 5 Years Randomised Controlled Trial." British Medical Journal 315: Treatment of High Blood Cholesterol in Adults). 2001. "Executive 786­90. Summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Cooper, R., C. Rotimi, S. Ataman, D. McGee, B. Osotimehin, S. Kadiri, Treatment of High Blood Cholesterol in Adults (Adult Treatment W. Muna, et al. 1997. "The Prevalence of Hypertension in Seven Panel III)." Journal of the American Medical Association 285 (19): Populations of West African Origin." American Journal of Public Health 2486­97. 87 (2): 160­68. Aspray, T. J., F. Mugusi, S. Rashid, D. Whiting, R. Edwards, K. G. Alberti, Cooper, R. S., C. N. Rotimi, J. S. Kaufman, W. F. Muna, and G. A. Mensah. and N. C. Unwin. 2000. "Rural and Urban Differences in Diabetes 1998. "Hypertension Treatment and Control in Sub-Saharan Africa: Prevalence in Tanzania: The Role of Obesity, Physical Inactivity and The Epidemiological Basis for Policy." British Medical Journal 316: Urban Living." Transactions of the Royal Society of Tropical Medicine 614­17. and Hygiene 94 (6): 637­44. Corrao, M. A., G. E. Guindon, N. Sharma, and D. F. Shokoohi. 2000. Baker, E. H., Y. B. Dong, G. A. Sagnella, M. Rothwell, A. K. Onipinla, N. D. Tobacco Control Country Profiles. Atlanta: American Cancer Society. Markandu, F. P. Cappuccio, et al. 1998. "Association of Hypertension Crowther, N. J., N. Cameron, J. Trusler, and I. P. Gray. 1998. "Association with T594M Mutation in Beta Subunit of Epithelial Sodium Channels Between Poor Glucose Tolerance and Rapid Post Natal Weight Gain in in Black People Resident in London." Lancet 351: 1388­92. Seven-Year-Old Children." Diabetologia 41: 1163­67. Barker, D. J. P. 1993. Fetal and Infant Origins of Adult Disease. London: Cusi, D., C. Barlassina, T. Azzani, G. Casari, L. Citterio, M. Devoto, British Medical Journal Publishing Group. N. Glorioso, et al. 1997. "Polymorphisms of Alpha-Adducin and Salt 262 | Krisela Steyn and Albertino Damasceno Sensitivity in Patients with Essential Hypertension." Lancet 349: Iyawe, V. I., A. D. Ighoroje, and H. O. Iyawe. 1996. "Changes in Blood 1353­57. Pressure and Serum Cholesterol Following Exercise Training in Damasceno, A., and A. Prista. 2001."Secular Trends in Cardiovascular Risk Nigerian Hypertensive Subjects." Journal of Human Hypertension Factors in Children and Youth from Maputo, Mozambique." Japanese 10 (7): 483­87. Journal of Cardiovascular Disease Prevention 36 (Suppl.): M024. Jain, S., X. Tang, C. S. Narayanan, Y. Agarwal, S. M. Peterson, C. D. Brown, Edwards, R., N. C. Unwin, F. Mugusi, D. Whiting, S. Rashid, J. Kissima, J. Ott, and A. Kumar. 2002. "Angiotensinogen Gene Polymorphism at T. J. Aspray, and K. G. Alberti. 2000. "Hypertension Prevalence and 217 Affects Basal Promoter Activity and Is Associated with Care in an Urban and Rural Area of Tanzania." Journal of Hypertension Hypertension in African-Americans." Journal of Biological Chemistry 18: 145­52. 277: 36889­96. Ekpo, E. B., O. Usofia, N. F. Eshiet, and J. J. Andy. 1992. "Demographic, Johns, T., R. L. Mahunnah, P. Samaya, L. Chapman, and T. Ticktin. 1999. Lifestyle and Anthropometric Correlates of Blood Pressure of Nigerian "Saponins and Phenolic Content in Plant Dietary Additives of a Urban Civil Servants, Factory and Plantation Workers." Journal of Traditional Subsistence Community, the Batemi of Ngorongoro Human Hypertension 6 (4): 275­80. District, Tanzania." Journal of Ethnopharmacology 66: 1­10. Ezenwaka, C. E., A. O. Akanji, B. O. Akanje, N. C. Unwin, and C. A. Joint Task Force. 1998. "Prevention of Coronary Heart Disease in Clinical Adejuwon. 1997. "The Prevalence of Insulin Resistance and Other Practice." Recommendations of the Second Joint Task Force of Cardiovascular Disease Risk Factors in Healthy Elderly Southwestern European and Other Societies on coronary prevention. European Heart Nigerians." Atherosclerosis 128 (2): 201­11. Journal 19 (10): 1434­1503. Fontaine, K. R., D. T. Redden, C. Wang, A. O. Westfall, and D. B. Allison. Kadiri, S., and B. L. Salako. 1997. "Cardiovascular Risk Factors in Middle 2003. "Years of Life Lost Due to Obesity." Journal of the American Aged Nigerians." East African Medical Journal 74 (5): 303­6. Medical Association 289: 187­93. Kaufman, J. S., E. E. Owoaje, C. N. Rotimi, and R. S. Cooper. 1999. "Blood Forrest, K. Y., C. H. Bunker, A. M. Kriska, F. A. Ukoli, S. L. Huston, and Pressure Change in Africa: Case Study from Nigeria." Human Biology N. Markovic. 2001. "Physical Activity and Cardiovascular Risk Factors 71 (4): 641­57. in a Developing Population." Medicine and Science in Sports and Kaufman, J. S., C. N. Rotimi, W. R. Brieger, M. A. Oladokum, S. Kadiri, B. Exercise 33 (9): 1598­1604. O. Osotimehin, and R. S. Cooper. 1996. "The Mortality Risk Associated Garnier, D., and E. Benefice. 2001. "Habitual Physical Activity of with Hypertension: Preliminary Results of a Prospective Study in Rural Senegalese Adolescent Girls under Different Working Conditions, As Nigeria." Journal of Human Hypertension 10 (7): 461­64. Assessed by a Questionnaire and Movement Registration." Annals of Kinabo, J. L., K. Kissawke, and J. Msuya. 1997. "Factors Influencing Birth Human Biology 28 (1): 79­97. Weight in Morogoro Municipality, Tanzania." South African Journal Gaziano, T. A., K. Steyn, and L. H. Opie. 2001. "Cost-Effectiveness Analysis Food Science and Nutrition 9: 3­8. of Hypertension Guidelines in South Africa." Abstract presented at the Knuiman, J. T., R. J. J. Hermus, and J. G. A. J. Hautvast. 1980. "Serum Total American Heart Association Meeting, Johannesburg, November and High-Density Lipoprotein (HDL) Cholesterol Concentrations in 10­13. Rural and Urban Boys from 16 Countries." Atherosclerosis 36: 529­37. Gibney, M. J., and P. G. Burstyn. 1980. "Milk, Serum Cholesterol and the Kruger, H. S., C. S. Venter, H. H. Vorster, and B. M. Margetts. 2002. Maasai. A Hypothesis." Atherosclerosis 35: 339­43. "Physical Inactivity Is the Major Determinant of Obesity in Black Global Youth Tobacco Survey Collaborative Group. 2002. "Tobacco Use Women in the North West Province, South Africa: The THUSA Study." among Youth: A Cross-Country Comparison." Tobacco Control 11: Nutrition 18: 422­27. 252­70. Lambert, E. V., W. Weitz, K. Charlton, M. I. Lambert, Z. Kukubeli, Grundy, S. M., R. Pasternak, P. Greenland, S. Smith Jr., and V. Fuster. 1999. L. Keytel, N. Temple, and A. Daniels. 2000. "Health and Fitness Survey "AHA/ACC Scientific Statement: Assessment of Cardiovascular Risk by in Adolescent School Children in the Western Cape: Relationship to Use of Multiple-Risk-Factor Assessment Equations: A Statement for Body Mass Index and Obesity." Paper presented at the biannual meet- Healthcare Professionals from the American Heart Association and the ing of the Nutrition Society and Association for Dietetics in Southern American College of Cardiology." Journal of the American College of African, "From Lab to Land," Durban, August. Cardiology 34 (4): 1348­59. Larson, N., R. Hutchinson, and E. Boerwinkle. 2000. "Lack of Association Haskell, W. L., A. S. Leon, and C. J. Caspersen. 1992. "Cardiovascular of 3 Functional Gene Variants with Hypertension in African Benefits and Assessment of Physical Activity and Physical Fitness in Americans." Hypertension 35: 1297­1300. Adults." Medicine and Science in Sports and Exercise 24 (Suppl.): Levitt, N. S., E. V. Lambert, D. Woods, C. N. Hales, R. Andrew, and J. R. S201­20. Seckl. 2000. "Impaired Glucose Tolerance and Elevated Blood Pressure Heini, A. F., G. Minghelli, E. Diaz, A. M. Prentice, and Y. Schutz. 1996."Free- in Low Birth Weight, Nonobese, Young South African Adults: Early Living Energy Expenditure Assessed by Two Different Methods in Rural Programming of Cortisol Axis." Journal of Clinical Endocrinology and Gambian Men." European Journal of Clinical Nutrition 50 (5): 184­89. Metabolism 85 (12): 4611­18. Hesse, I. F., and I. Nuama. 1997. "Pattern of Out-Patient Drug Treatment Levitt, N. S., K. Steyn, T. De Wet, C. Morrell, P. R. Edwards, G. T. H. Ellison, of Hypertension in Korle-Bu Teaching Hospital, Accra." West African and N. Cameron. 1999. "An Inverse Relation between Blood Pressure Journal of Medicine 16 (3): 133­38. and Birth Weight among 5-Year-Old Children from Soweto, South Holman, H., and K. Lorig. 2000. "Patients as Partners in Managing Africa (Birth-to-Ten Study)." Journal of Epidemiology and Community Chronic Disease." British Medical Journal 320: 526­27. Health 53: 264­68. Howson C. P., K. S. Reddy, T. J. Ryan, and J. R. Bale, eds. 1998. Control of Levitt, N. S., K. Steyn, E. V. Lambert, G. Reagon, C. J. Lombard, J. M. Cardiovascular Disease in Developing Countries: Research, Development, Fourie, K. Rossouw, and M. Hoffman. 1999. "Modifiable Risk Factors and Institutional Strengthening. Washington, DC: National Academy for Type 2 Diabetes Mellitus in a Peri-Urban Community in South Press. Africa." Diabetic Medicine 16: 946­50. Hunter, J. M., B. T. Sparks, J. Mufunda, C. T. Musabayane, H. V. Sparks, and Longo-Mbenza, B., R. Ngiyulu, M. Bayekula, E. K. Vita, F. B. Nkiabungu, K. Mohamed. 2000. "Economic Development and Women's Blood K. V. Seghers, E. L. Luila, F. M. Mandundu, and M. Manzanza. 1999. Pressure: Field Evidence from Rural Mashonaland, Zimbabwe." Social "Low Birth Weight and Risk of Hypertension in African School Science and Medicine 50 (6): 773­95. Children." Journal of Cardiovascular Risk 6: 311­14. Lifestyle and Related Risk Factors for Chronic Diseases | 263 Margetts, B. M., M. G. M. Rowlands, F. A. Foord, A. M. Cruddas, T. J. Cole, Determinants of High Blood Pressure in a Group of Urban Nigerians." and D. J. P. Barker. 1991. "The Relation of Maternal Weight to the Journal of Human Hypertension 14(4): 249­57. Blood Pressure of Gambian Children." International Journal of Pauletto, P., M. Puato, M. G. Caroli, E. Casiglia, A. E. Munhambo, Epidemiology 20: 938­43. G. Cazzolato, G. Bittolo Bon, M. T. Angeli, C. Galli, and A. C. Pessina. Marks, A. S., K. Steyn, and E. Ratheb. 2001. "Tobacco Use by Black Women 1996. "Blood Pressure and Atherogenic Lipoprotein Profiles of Fish- in Cape Town." MRC Policy Brief 3. Parowvallei: Medical Research Diet and Vegetarian Villagers in Tanzania: The Lugalawa Study." Lancet Council. 348 (9030): 784­88. Mbanya, J. C .N., E. M. Minkoulou, J. N. Salah, and B. Balkau. 1998. "The Pavan, L., E. Casiglia, P. Pauletto, S. L. Batista, G. Ginocchio, M. M. Prevalence of Hypertension in Rural and Urban Cameroon." Kwankam, R. Biasin, et al. 1997. "Blood Pressure, Serum Cholesterol International Journal of Epidemiology 27: 181­85. and Nutritional State in Tanzania and in the Amazon: Comparison McBride, P., J. Einerson, P. Hanson, and K. Heindel. 1992."Exercise and the with an Italian Population." Journal of Hypertension 15 (10): 1083­90. Primary Prevention of Coronary Heart Disease." Medicine, Exercise, Peeters, A., J. J. Barendregt, F. Willekens, J. P. Mackenbach, A. A. Mamun, Nutrition, and Health 1: 5­15. L. Bonneux, for NEDCOM (Netherlands Epidemiology and McCormick, J., and M. Elmore-Meegan. 1992. "Maasai Diet." Lancet 340: Demography Compression of Morbidity Research Group). 2003. 1042­43. "Obesity in Adulthood and Its Consequences for Life Expectancy: A Life-Table Analysis." Annals of Internal Medicine 138: 24­32. Milne, F. J., Y. Veriava, S. H. James, and C. Isaacson. 1989. "Aetiology and Pathogenesis of Malignant Hypertension in Black South Africans--A Popkin, B. M. 2001a. "Nutrition in Transition: The Changing Global Review." South African Medical Journal 76 (Suppl.): 22­23. Nutrition Challenge." Asia Pacific Journal of Clinical Nutrition 10: S13­18. Mollentze, W. F., A. J. Moore, G. Joubert, K. Steyn, G. M. Oosthuizen, and D. J. Weich. 1995. "Coronary Heart Disease Risk Factors in a Rural and ------. 2001b. "The Nutrition Transition and Obesity in the Developing Urban Orange Free State Black Population." South African Medical World." Journal of Nutrition 131 (3): 871S­73S. Journal 85 (2): 90­96. Poulter, N., K. T. Khaw, B. E. Hopwood, M. Mugambi, W. S. Peart, and Mtabaji, J. P., Y. Moriguchi, Y. Nara, S. Mizushima, M. Mano, and Y. P. S. Sever. 1985. Determinants of Blood Pressure Changes Due to Yamori. 1992." Ethnic Differences in Salt Sensitivity: Genetic or Urbanization: A Longitudinal Study. Journal of Hypertension 3 Environmental Factors." Clinical and Experimental Pharmacology and (Suppl.): S375­77. Physiology 20 (Suppl.): 65­67. Poulter, N., K. T. Khaw, B. E. Hopwood, M. Mugambi, W. S. Peart, G. Rose, Mtabaji, J. P., Y. Nara, and Y. Yamori. 1990."The Cardiac Study in Tanzania: and P. S. Sever. 1990. "The Kenyan Luo Migration Study: Observations Salt Intake in the Causation and Treatment of Hypertension." Journal on the Initiation of a Rise in Blood Pressure." British Medical Journal of Human Hypertension 4 (32): 80­81. 300: 967­72. Mufunda, J., L. J. Scott, J. Chifamba, J. Matenga, B. Sparks, R. Cooper, and Puoane, T., K. Steyn, D. Bradshaw, R. Laubscher, J. Fourie, V. Lambert, H. Sparks. 2000. "Correlates of Blood Pressure in an Urban and N. Mbananga. 2002. "Obesity in South Africa: The South Zimbabwean Population and Comparison to Other Populations of African Demographic and Health Survey." Obesity Research 10 (10): African Origin." Journal of Human Hypertension 14 (1) : 65­73. 1038­48. Mvo, Z., J. Dick, and K. Steyn. 1999. "Perceptions of Overweight African Rotimi, C. N., R. S. Cooper, S. L. Ataman, B. Osotimehin, S. Kadiri, W. Women about Acceptable Body Size and Children." Curationis 22 (2): Muna, S. Kingue, H. Fraser, and D. McGee. 1995. "Distribution of 27­31. Anthropometric Variables and the Prevalence of Obesity in Ndlovo, P. P., and S. D. Roos. 1999. "Perceptions of Black Women of Populations of West African Origin: The International Collaborative Obesity as a Health Risk." Curationis 22 (2): 47­55. Study on Hypertension in Blacks (ICSHIB)." Obesity Research 2 (Suppl.): 95S­105S. Nissinen, A., S. Bothig, H. Granroth, and A. D. Lopez. 1988. "Hypertension in Developing Countries." World Health Statistics Quarterly 41 (3­4): Rotimi, C. N., R. S. Cooper, G. Cao, O. Ogunbiyi, M. Ladipo, E. Owoaje, 141­54. and R. Ward. 1999. "Maximum-Likelihood Generalized Heritability Estimate for Blood Pressure in Nigerian Families." Hypertension 33 (3): Njelekela, M., H. Negishi, Y. Nara, M. Tomohiro, S. Kuga, T. Noguchi, 874­78. T. Kanda, et al. 2001. "Cardiovascular Risk Factors in Tanzania: A Revisit." Acta Tropica 79 (3): 231­39. Salako, B. L., S. Kadiri, F. A. Fehintola, and O. O. Akinkugbe. 1999. "The Effect of Anti-Hypertensive Therapy on Urinary Albumin Excretion in Nkeh, B., N. J. Samani, D. Badenhorst, E. Libhaber, P. Sareli, G. R. Norton, Nigerian Hypertensives." West African Journal of Medicine 18 (3): and A. J. Woodiwiss. 2003. "T594M Variant of the Epithelial Sodium 170­74. Channel Beta-Subunit Gene and Hypertension in Individuals of African Ancestry in South Africa." American Journal of Hypertension 16 Saloojee, Y. 1995. "Price and Income Elasticity of Demand for Cigarettes in (10): 847­52. South Africa in Tobacco and Health." In Tobacco and Health, ed. K. Shlama. New York: Plenum Press. Obel, A. O., and D. K. Koech. 1991. "Potassium Supplementation versus Bendro Fluazide in Mildly to Moderately Hypertensive Kenyans." ------. 2000. "Regional Summary for the African Region." In Tobacco Journal of Cardiovascular Pharmacology 17 (3): 504­7. Control Country Profiles, ed. M. A. Corrao, G. E. Guindon, N. Sharma, and D. F. Shokoohi. Atlanta: American Cancer Society. Odendaal, H. J., D. L. Van Schie, and R. M. De Jeu. 2001. "Adverse Effects of Maternal Cigarette Smoking on Preterm Labour and Abruptio Seedat, Y. K. 1996. "Is the Pathogenesis of Hypertension Different in Black Placentae." International Journal Gynecology and Obstetrics 74: 287­88. Patients?" Journal of Human Hypertension 3 (Suppl.): S35­37. Oelofse, A., P. Jooste, K. Steyn, C. J. Badenhorst, L. T. Bourne, and J. Fourie. ------. 1999. "Hypertension in Black South Africans." Journal of Human 1996. "The Lipid and Lipoprotein Profile of the Urban Black Hypertension 13 (2): 96­103. South African Community of the Cape Peninsula--the BRISK Study." Seftel, H., M. S. Asvat, B. I. Joffe, F. J. Raal, V. R. Panz, W. J. H. Vermaak, South African Medical Journal 86: 166­69. M. E. Loock, et al. 1993. "Selected Risk Factors for Coronary Heart Olatunbosun, S. T., J. S. Kaufman, R. S. Cooper, and A. F. Bella. 2000. Disease in Male Scholars from the Major South African Population "Hypertension in a Black Population: Prevalence and Biosocial Groups." South African Medical Journal 83: 891­97. 264 | Krisela Steyn and Albertino Damasceno Seftel, H., F. J. Raal, and B. I. Joffe. 1995. "Dyslipidaemia in South Africa." Swai, A. B., D. G. McLarty, H. M. Kitange, P. M. Kilima, S. Tatalla, N. Keen, In Chronic Diseases of Lifestyle in South Africa, ed. J. M. Fourie, and L. M. Chuwa, and K. G. Alberti. 1993. "Low Prevalence of Risk Factors K. Steyn. MRC Technical Report. Parowvallei, Medical Research for Coronary Heart Disease in Rural Tanzania." International Journal of Council. Epidemiology 22 (4) : 651­59. Siffert, W., D. Rosskopf, G. Siffert, S. Busch, A. Moritz, R. Erbel, A. M. Swartz, L., and J. Dick. 2002. "Managing Chronic Conditions in Less Sharma, et al. 1998. "Association of a Human G-Protein Beta3 Subunit Developed Countries." British Medical Journal 325: 914­15. Variant with Hypertension." Nature Genetics 18: 45­48. Teklu, T. 1996. "Food Demand Studies in Sub-Saharan Africa: A Survey of Sobngwi, E., J. C. N. Mbanya, N. C. Unwin, A. P. Kengne, L. Fezeu, E. M. Empirical Evidence." Food Policy 21: 479­96. Minkoulou, T. J. Aspray, and K. G. Alberti. 2002. "Physical Activity and Tiago, A. D., N. J. Samani, G. P. Candy, R. Brooksbank, E. N. Libhaber, Its Relationship with Obesity, Hypertension and Diabetes in Urban P. Sareli, A. J. Woodiwiss, and G. R. Norton. 2002. "Angiotensinogen and Rural Cameroon." International Journal of Obesity and Related Gene Promoter Region Variant Modifies Body Size-Ambulatory Blood Metabolic Disorders 26 (7): 1009­16. Pressure Relations in Hypertension." Circulation 106: 1483­87. Sobngwi, E., J. C. N. Mbanya, N. C. Unwin, J. A. Terrence, and K. G. M. M. Treloar, C., J. Porteous, F. Hassan, N. Kasniyah, M. Lakshmanudu, M. Alberti. 2001. "Development and Validation of a Questionnaire for Sama, M. Sha'bani, and F. Heller. 1999. "The Cross Cultural Context of the Assessment of Physical Activity in Epidemiological Studies in Obesity: An INCLEN Multicentre Collaborative Study." Health and Sub-Saharan Africa." International Journal of Epidemiology 30: Place 5: 279­86. 1361­68. Unwin, N., P. Setel, S. Rashid, F. Mugusi, J. C. Mbanya, H. Kitange, Solomon, C. G., and J. E. Manson. 1997. "Obesity and Mortality: A Review L. Hayes, R. Edwards, T. Aspray, and K. G. Alberti. 2001. of the Epidemiologic Data." American Journal of Clinical Nutrition 66 "Noncommunicable Diseases in Sub-Saharan Africa: Where Do They (Suppl.): 1044S­50S. Feature in the Health Research Agenda?" Bulletin of the World Health Steyn, K., L. T. Bourne, P. L. Jooste, J. M. Fourie, C. J. Lombard, and Organization 79 (10): 947­53. D. Yach. 1994. "Smoking in the African Community of the Cape Van der Sande, M. A. B., P. J. M. Milligan, O. A. Nyan, J. T. Rowley, W. A. S. Peninsula." East African Medical Journal 71: 784­89. Banya, S. M. Ceesay, W. M. Dolmans, T. Thien, K. P. W. J. McAdam, and Steyn, K., D. Bradshaw, R. Norman, R. Laubscher, and Y. Saloojee. 2002. G. E. L. Walvaren. 2000. "Blood Pressure Patterns and Cardiovascular "Tobacco Use in South Africans During 1998: The First Demographic Risk Factors in Rural and Urban Gambian Communities." Journal of and Health Survey." Journal of Cardiovascular Risk 9: 161­70. Human Hypertension 14 : 489­96. Steyn, K., T. de Wet, L. Richter, N. Cameron, N. S. Levitt, and C. Morrell. Van der Sande, M. A. B., G. E. L. Walvaren, P. J. M. Milligan, W. A. S. Banya, 2000. "Cardiovascular Disease Risk Factors in Five-Year-Old Urban S. M. Ceesay, O. A. Nyan, and K. P. W. J. McAdam. 2001. "Family South African Children--The Birth to Ten Study." South African History: An Opportunity for Early Interventions and Improved Medical Journal 90 (7): 719­26. Control of Hypertension, Obesity and Diabetes." Bulletin of the World Steyn, K., J. M. Fourie, C. J. Lombard, J. Katzenellenbogen, L. Bourne, and Health Organization 79: 321­28. P. Jooste. 1996. "Hypertension in the Black Community of the Cape Veriava, Y., E. du Toit, C. G. Lawley, F. J. Milne, and S. G. Reinach. 1990. Peninsula, South Africa." East African Medical Journal 11: 758­63. "Hypertension as a Cause of End-Stage Renal Failure in South Africa." Steyn, K., J. M. Fourie, and J. Sheppard. 1998. "The Lipid Profile of South Journal of Human Hypertension 4 (4): 379­83. Africans Attending General Practitioners: The Cholesterol Monitor." Wagner, E. G. 2000. "The Role of Patient Care Teams in Chronic Care South African Medical Journal 88: 1569­74. Management." British Medical Journal 320: 569­671. Steyn, K., T. Gaziano, D. Bradshaw, R. Laubscher, and J. M. Fourie. 2001. Walker, A. R., and B. F. Walker. 1978. "High High-Density-Lipoprotein "Hypertension in South African Adults: Results from the Demographic Cholesterol in African Children and Adults in a Population Free of and Health Survey." Journal of Hypertension 19: 1717­25. Coronary Heart Disease." British Medical Journal 6148 (2): 1336­37. Steyn, K., Y. P. Goldberg, M. J. Kotze, M. Steyn, A. S. Swanepoel, J. M. Walker, A. R., B. F. Walker, B. Manetsi, N. G. Tsotetsi, and A. J. Walker. 1990. Fourie, G. A. Coetzee, and D. R. van der Westhuyszen. 1996. "Obesity in Black Women in Soweto, South Africa: Minimal Effects on "Estimation of the Prevalence of Familial Hypercholesterolaemia in a Hypertension, Hyperlipidaemia and Hyperglycemia." Journal of the Rural Afrikaner Community by Direct Screening for Three Afrikaner Royal Society of Health 110 (3): 101­3. Founder Low-Density Lipoprotein Receptor Gene Mutations." Human Wareham, J. 2001. "Commentary: Measuring Physical Activity in Sub- Genetics 98 (4): 479­84. Saharan Africa." International Journal of Epidemiology 30: 1369­70. Steyn, K., D. Yach, I. Stander, and J. M. Fourie. 1997. "Smoking in Urban WHO (World Health Organization). 2000. Obesity: Preventing and Pregnant Women in South Africa." South African Medical Journal 87: Managing the Global Epidemic. WHO Technical Report Series 894. 460­63. Geneva: WHO. Steyn, N. P., M. Senekal, S. Brits, and J. H. Nel. 2000. "Urban and Rural Differences in Dietary Intake, Weight Status and Nutrition ------. 2002. Innovative Care for Chronic Conditions: Building Blocks for Knowledge of Black Female Students." Asia Pacific Journal of Clinical Action. WHO Global Report. Geneva: WHO. Nutrition 9 (1): 11­16. ------. 2003 (updated 2004/5). "WHO Framework Convention on St-Onge, M. P., E. R. Farnworth, and P. J. Jones. 2000. "Consumption of Tobacco Control." http://www.who.int/tobacco/framework/WHO_ Fermented and Nonfermented Dairy Products: Effects on Cholesterol FCTC_english.pdf (accessed January 31, 2006). Concentrations and Metabolism." American Journal of Clinical Yach, D. 2002. "Unleashing the Power of Prevention to Achieve Global Nutrition 71 (3): 674­81. Health-Gains." Lancet 360: 1343­44. Svetkey, L. P., P. Z. Timmons, O. Emovon, N. B. Anderson, L. Preis, and Y. T. Zhao, Y. Y., J. Zhou, C. S. Narayanan, Y. Cui, and A. Kumar. 1999. "Role of Chen. 1996. "Association of Hypertension with Beta2- and Alpha2c10- C/A Polymorphism at 20 on the Expression of Human Adrenergic Receptor Genotype." Hypertension 27 (6): 1210­15. Angiotensinogen Gene." Hypertension 33: 108­15. Lifestyle and Related Risk Factors for Chronic Diseases | 265 Chapter 19 Diabetes Mellitus Jean-Claude Mbanya and Kaushik Ramiaya The global burden of disease study of the World Health changes in risk-factor levels, such as tobacco smoking, Organization (WHO) estimated that about 177 million obesity, and physical inactivity (Hunter et al. 2000; Kaufman people in the world had diabetes in the year 2000 (WHO et al. 1999; Pavan et al. 1997). Countries of Sub-Saharan 2003). In the second edition of the International Diabetes Africa are in various stages of the epidemiological transi- Federation's Diabetes Atlas it is estimated that 194 million tion with a multiple burden of diseases. people had diabetes in the year 2003, and about two-thirds of The available evidence suggests that noncommunicable these people lived in developing countries (IDF 2003). In diseases currently contribute substantially to the burden of 1901 Albert Cook, a medical missionary in Uganda, reported mortality and morbidity in adults. Age-specific levels of dia- that "diabetes is rather uncommon and very fatal" (Cook betes and hypertension in many urban areas of Sub-Saharan 1901). Over the next 50 to 60 years diabetes continued to Africa are as high as, or higher than, those in most Western be regarded as rare in Sub-Saharan Africa. Communicable European countries (Aspray et al. 2000; Edwards et al. 2000; diseases still make up the greatest disease burden, but by Mollentze et al. 1995). In a demographic surveillance system 2020, noncommunicable diseases, including hypertension in Tanzania they account for between one in six and one in and diabetes, will outstrip communicable diseases as a cause three adult deaths (Kitange et al. 1996; Setel et al. 2000; of death (Murray and Lopez 1997). Even allowing for the Walker et al. 2000), with age-specific death rates from non- uncertainties of predicting future disease patterns posed by specific, noncommunicable diseases being as high or higher the unfolding of the human immunodeficiency virus (HIV) than in developed countries (Unwin et al. 1999). epidemic in Sub-Saharan Africa, it is clear that the relative Diabetes mellitus can be classified into four principal importance of noncommunicable diseases will increase types (WHO 1999). This includes type 1 diabetes, type 2 (Panz and Joffe 1999). This situation is a result of demo- diabetes, other specific types of diabetes, and gestational dia- graphic change (populations with older age structures), betes mellitus. The most common types of diabetes seen in increasing urbanization (WHO 1998), and associated Sub-Saharan Africa are type 2 and type 1 diabetes mellitus. 267 This chapter focuses on the published data on the burden of information on the burden of diabetes for their Sub- type 1 and type 2 diabetes in Sub-Saharan Africa. Although Saharan Africa country or subregion. The Medline search type 1 diabetes is not caused by the adverse effects of was undertaken for diabetes prevalence and for each com- lifestyle, as type 2 can be, the chronic complications of both plication: retinopathy, neuropathy, nephropathy, and so on. type 1 and type 2 diabetes on the eyes, cardiovascular In the absence of data from any country, data were extrapo- system, nerves, and kidneys are similar. lated from the socioeconomically, ethnically, and geograph- ically most similar country. The data obtained were from SOURCES OF DATA prevalence studies, hospital-based studies, registry reports, hospital statistics, government estimates, and the like. Some The data search was limited to studies published after 1979, of the specific data sources are mentioned in table 19.1. The because data collected before 1980 may no longer reflect the sources of data for the 2003 estimates of diabetes mellitus current prevalence of diabetes. The Medline database and and IGT are listed in table 19.2. the Internet were used for the literature search, but also There is still a dearth of published studies describing the diabetes researchers and clinicians were asked to provide burden of diabetes in Sub-Saharan Africa. The prevalence Table 19.1 Data Sources for the Prevalence of Type 2 Diabetes and IGT, by Year of Study Prevalence (%) Country, locality Data sources Year Urban-rural Age group (yrs) N DM IGT Tanzania Ahren and Corrigan 1984 U and R 20 3,145 0.7 -- Kahalanga R 996 0.5 -- Ndolage R 1,141 2.5 -- Mwanza U 1,008 1.9 -- Mali Fisch et al. 1987 R 15 7,472 0.9 -- Togo Teuscher et al. 1987 R 1 1,381 0.0 -- Nigeria Ohwovoriole, Kuti, and Kabiawu 1988 U -- 1,627 1.7 -- Tanzania McLarty et al. 1989 R 15 6,097 0.9 7.8 1.1 8.4a South Africa Cape Town Levitt et al. 1993 U 30 729 6.3 5.9 8.0a 7.0a Durban Omar et al. 1993 U 15 479 4.2 6.9 5.3a 7.7a Mangaung, OFS Mollentze et al. 1995 U 25 758 6.0a 12.2a Qwa-Qwa, OFS R 853 4.8a 10.7a Mauritania Ducorps et al. 1996 U and R 17 744 1.88 -- Cameroon Mbanya et al. 1997 U and R 24­74 1,767 1.1 2.7 Yaoundé U 1,048 1.3 1.8 Evodoula R 719 0.8 3.9 Nigeria Cooper et al. 1997 R 25­74 247 2.8 -- Tanzania Aspray et al. 2000 U and R 15 Dar es Salaam U 770 -- -- Kilimanjaro R 928 -- -- Ghana Amoah, Owusu, and Adjei 2002 U 25 4,733 6.3 10.7a 6.4a Source: Authors, from data sources in table. Note: The studies were published using the WHO 1980, 1985, or 1999 criteria. DM diabetes mellitus (type 2 diabetes); IGT impaired glucose tolerance; -- not available; OFS Orange Free State. a. Age-adjusted rates. 268 | Jean-Claude Mbanya and Kaushik Ramiaya Table 19.2 Data Sources for Prevalence Estimates of Diabetes Mellitus and IGT, by Country, 2003 Screening Diagnostic Sample Country Data sources method criteria size Age Angolaa McLarty et al. 1989 and Aspray et al. 2000 (Tanzania) OGTT/FBG WHO, 1985 and 1999 7,781 15 Beninb Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Botswanac Omar et al. 1993 and Levitt et al. 1993 (South Africa) OGTT WHO, 1985 1,208 15 Burkina Fasob Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Burundia McLarty et al. 1989 and Aspray et al. 2000 (Tanzania) OGTT/FBG WHO, 1985 and 1999 7,781 15 Cameroon Mbanya et al. 1997 (Cameroon) OGTT/FBG WHO, 1985 1,767 24­74 Cape Verdeb Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Central African Republica Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Chad Elbagir et al. 1996 (Sudan) 2BG WHO, 1985 1,284 25­84 Comoros McLarty et al. 1989 and Aspray et al. 2000 (Tanzania) OGTT/FBG WHO, 1985 and 1999 7,781 15 Congo, Dem. Rep. ofa McLarty et al. 1989 and Aspray et al. 2000 (Tanzania) OGTT/FBG WHO, 1985 and 1999 7,781 15 Congo, Rep. ofa Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Côte d'Ivoire Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Djibouti Elbagir et al. 1996 (Sudan) 2BG WHO, 1985 1,284 25­84 Equatorial Guineab Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Eritreaa McLarty et al. 1989 and Aspray et al. 2000 (Tanzania) OGTT/FBG WHO, 1985 and 1999 7,781 15 Ethiopia McLarty et al. 1989 and Aspray et al. 2000 (Tanzania) OGTT/FBG WHO, 1985 and 1999 7,781 15 Gabonb Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Gambia, Theb Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Ghana Amoah, Owusu, and Adjei 2002 (Ghana) OGTT WHO, 1999 4,733 25 Guineab Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Guinea-Bissaub Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Kenyaa McLarty et al. 1989 and Aspray et al. 2000 (Tanzania) OGTT/FBG WHO, 1985 and 1999 7,781 15 Lesothoc Omar et al. 1993 and Levitt et al. 1993 (South Africa) OGTT WHO, 1985 1,208 15 Liberiab Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Madagascara McLarty et al. 1989 and Aspray et al. 2000 (Tanzania) OGTT/FBG WHO, 1985 and 1999 7,781 15 Malawia McLarty et al. 1989 and Aspray et al. 2000 (Tanzania) OGTT/FBG WHO, 1985 and 1999 7,781 15 Malib Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Mauritania Elbagir et al. 1996 (Sudan) 2BG WHO, 1985 1,284 25­84 Mozambiquea McLarty et al. 1989 and Aspray et al. 2000 (Tanzania) OGTT/FBG WHO, 1985 and 1999 7,781 15 Namibiac Omar et al. 1993 and Levitt et al. 1993 (South Africa) OGTT WHO, 1985 1,208 15 Nigerb Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) (Continues on the following page.) Diabetes Mellitus | 269 Table 19.2 (Continued) Screening Diagnostic Sample Country Data sources method criteria size Age Nigeriab Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Réunion Dowse et al. 1990 (Mauritius) OGTT WHO, 1985 4,929 25­74 Rwandaa McLarty et al. 1989 and Aspray et al. 2000 (Tanzania) OGTT/FBG WHO, 1985 and 1999 7,781 15 São Tomé and Principeb Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Senegalb Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Seychelles Dowse et al. 1990 (Mauritius) OGTT WHO, 1985 4,929 25­74 Sierra Leoneb Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Somaliaa McLarty et al. 1989 and Aspray et al. 2000 (Tanzania) OGTT/FBG WHO, 1985 and 1999 7,781 15 South Africac Omar et al. 1993 and Levitt et al. 1993 (South Africa) OGTT WHO, 1985 1,208 15 Swazilandc Omar et al. 1993 and Levitt et al. 1993 (South Africa) OGTT WHO, 1985 1,208 15 Tanzaniaa McLarty et al. 1989 and Aspray et al. 2000 (Tanzania) OGTT/FBG WHO, 1985 and 1999 7,781 15 Togob Mbanya et al. 1997 (Cameroon) and Amoah, Owusu, OGTT WHO, 1985 and 1999 6,500 24 and Adjei 2002 (Ghana) Ugandaa McLarty et al. 1989 and Aspray et al. 2000 (Tanzania) OGTT/FBG WHO, 1985 and 1999 7,781 15 Western Sahara Elbagir et al. 1996 (Sudan) 2BG WHO, 1985 1,284 25­84 Zambiaa McLarty et al. 1989 and Aspray et al. 2000 (Tanzania) OGTT/FBG WHO, 1985 and 1999 7,781 15 Zimbabwea Omar et al. 1993 and Levitt et al. 1993 (South Africa) OGTT WHO, 1985 1,208 15 Source: Adapted with permission from IDF 2003. Note: See note for table 19.1; OGTT oral glucose tolerance test; FBG fasting blood glucose; 2BG 2 hours blood glucose. a. The prevalence was calculated after the combination of the data of the two studies, notwithstanding the different criteria. IGT figures were calculated from the McLarty et al. 1989 data, as the Aspray et al. 2000 study used only the FBG criteria. b. The prevalence was calculated as the average of the two studies, as their sample sizes differed considerably. c. The prevalence was calculated after the combination of the data of the two studies. IGT figures were based only on the study of Omar et al. 1993. rates of diabetes and its complications (table 19.3) have been Type 1 diabetes results from autoimmune destruction of drawn from the few country data available as applied to the the pancreatic beta cells, causing the loss of insulin produc- population distribution of that country or a similar country. tion. Children are usually affected by this type of diabetes, Clinic-based studies have serious limitations, so their gener- although it occurs at all ages and the clinical presentation alizability is limited. Therefore the data presented here are can vary with age. Patients with this type of diabetes require only general indicators of diabetes frequency and should be insulin for survival. interpreted with caution. As new and better epidemiological Type 2 diabetes is characterized by insulin resistance and data become available, it will be possible to have actual rates abnormal insulin secretion, either of which may predomi- of diabetes in Sub-Saharan Africa. nate but both of which are usually present. The specific reasons for the development of these abnormalities are largely unknown. Type 2 is the most common type of EPIDEMIOLOGY OF DIABETES diabetes. Type 2 diabetes can remain asymptomatic for many years, and the diagnosis is often made from associated Diabetes mellitus is a chronic metabolic disease character- complications or incidentally through an abnormal blood ized by hyperglycemia resulting from defects in insulin or urine glucose test. secretion, insulin action, or both. Uncontrolled chronic Other specific types of diabetes include those due to hyperglycemia results in long-term damage, particular genetic disorders, infections, diseases of the exocrine pan- dysfunction, and failure of the eyes, heart, blood vessels, creas, endocrinopathies, and drugs. This last type of diabetes nerves, and kidneys. is relatively uncommon. 270 | Jean-Claude Mbanya and Kaushik Ramiaya Table 19.3 Data Sources for the Prevalence of Diabetes Complications, by Disease and Year Prevalence Year Data source Country Study type (%) Retinopathy 1988 Rolfe Zambia Clinic (secondary care) 34 1993 Lester Ethiopia Clinic (secondary care) 13 1995 Gill, Huddle, and Rolfe South Africa Clinic (secondary care) 52 1995 Moukouri et al. Cameroon Clinic (secondary care) 37 1996 Drabo, Kabore, and Lengani Burkina Faso Clinic (secondary care) 16 1997 Kalk et al. South Africa Clinic (secondary care) 37 1997 Levitt et al. South Africa Clinic (primary care) 55 1997 Rahlenbeck and Gebre-Yohannes Ethiopia Clinic (secondary care) 36 1999 Sobngwi et al. Cameroon Clinic (secondary care) 37 Nephropathy 1996 Drabo, Kabore, and Lengani Burkina Faso Clinic (secondary care) 25 1997 Levitt et al. South Africa Clinic (primary care) 37 1997 Rahlenbeck and Gebre-Yohannes Ethiopia Clinic (secondary care) 33 1999 Sobngwi et al. Cameroon Clinic (secondary care) 46a Neuropathy 1991 Lester Ethiopia Clinic (secondary care) 36 1995 Gill, Huddle, and Rolfe South Africa Clinic (secondary care) 42 1997 Levitt et al. South Africa Clinic (primary care) 28 1988 Rolfe Zambia Clinic (secondary care) 31 Vascular disease, lower limbs 1994 Niang et al. Senegal Clinic (secondary care) 28 1997 Levitt et al. South Africa Clinic (primary care) -- Coronary artery disease 1996 Drabo, Kabore, and Lengani Burkina Faso Clinic (secondary care) 28 1996 Nambuya et al. Uganda Clinic (secondary care) 5 1988 Rolfe Zambia Clinic (secondary care) 1 Source: Authors. Note: -- not available. a. Microalbuminuria. Gestational diabetes mellitus (GDM) is defined as any tolerance test. Impaired fasting glycemia (IFG) is an degree of glucose intolerance with onset or first recognition elevated nondiabetic fasting blood glucose level. Both IGT during pregnancy. The definition applies whether insulin and IFG are transitional stages in the development of or only diet modification is used for treatment and whether type 2 diabetes. the condition persists after pregnancy. It does not exclude the possibility that unrecognized glucose intolerance may have antedated or begun concomitantly with the pregnancy. PREVALENCE AND INCIDENCE OF TYPE 1 Approximately 7 percent of all pregnancies are complicated DIABETES by GDM. The prevalence may range from 1 to 14 percent of all pregnancies, depending on the population studied and There is a dearth of published studies describing the inci- the diagnostic tests employed. dence and prevalence of type 1 diabetes in Sub-Saharan Impaired glucose tolerance (IGT) is asymptomatic, and Africa. Type 1 diabetes is considerably rarer than type 2 its diagnosis is confirmed by an elevated nondiabetic level disease, and large populations need to be surveyed. Also, of blood glucose two hours after a 75 gram oral glucose to assess incidence, the population surveyed should be Diabetes Mellitus | 271 accurately known, and this is in itself difficult, as complete negative associations with some alleles (Garcia-Pacheco censuses in Africa are rare and migration in and out of study et al. 1992; Chauffert et al. 1995). areas common. Elamin and colleagues in the Sudan in 1992 reported a survey of nearly 43,000 schoolchildren (age 7 to Immunological Factors 11 years) and found a prevalence rate of 0.95 per 1,000 (Elamin et al. 1992). This rate is comparable to a reported The main markers of immune islet cell attack are islet cell prevalence rate of 0.3 per 1,000 in Nigeria (Afoke et al. antibodies (ICA) and glutamic acid decarboxylase antibodies 1992). The reported incidence is 10.1 per 100,000 children (anti-GAD). These substances are found in most Caucasian per year in Sudan (Elamin et al. 1992) and 1.5 per 100,000 type 1 diabetic patients at diagnosis, but levels gradually per year in Tanzania (Swai, Lutale, and McLarty 1993). The decline with time. Interpretation of ICA and anti-GAD discrepancy between the Sudanese and Tanzanian studies levels in type 1 diabetes is dependent on duration of disease, may be explained by ethnic differences, and perhaps prob- and this may explain the variable results found in the limit- lems related to the design of the studies. ed African studies so far carried out. McLarty, Kinabo, and The question of whether type 1 diabetes is truly rarer in Swai (1990) found that the prevalence of ICA antibodies Africa than elsewhere remains unsettled, and more detailed was only 8 to 11 percent in newly diagnosed Tanzanian surveys are needed. Nonetheless, it emerges from careful patients. In South Africa, Motala, Omar, and Pirie (2000) clinic studies that the behavior of type 1 diabetes is different found that 44 percent of blacks with newly diagnosed type 1 in Sub-Saharan Africa from that in the rest of the world. diabetes were positive for GAD antibody. It appears from Studies indicate that the age of onset in South Africa and these preliminary results that the genetic susceptibility and Ethiopia is later than elsewhere (Kalk, Huddle, and Raal risk factors for type 1 diabetes in Sub-Saharan Africa may be 1993; Lester 1984), and the peak age of onset of type 1 dia- different from those in the Western world. It can be specu- betes in Sub-Saharan Africa is a decade later than in the lated that non-autoimmune factors are the major determi- West (Afoke et al. 1992; Kalk, Huddle, and Raal 1993). In nants of type 1 diabetes in Sub-Saharan Africa. addition it afflicts more females than males. In South Africa it has been reported that the peak age of onset was about Environmental Factors 13 years in the white South Africans (similar to Europeans) but about 23 years in the black South Africans (Kalk, An environmental "trigger" factor for the onset of type 1 Huddle, and Raal 1993). The reasons for this difference are diabetes has long been sought. Its existence is supported by obscure, although it has been suggested that prolonged the well-known seasonality of presentation in Europe, and breastfeeding, which is common in Africa, may be reducing viral infection (perhaps of the coxsackievirus group) is con- the incidence and delaying the onset of type 1 diabetes. Early sidered a likely candidate. A seasonality of type 1 diabetes introduction of cow's milk protein does seem to be a risk has been reported in Tanzania (with most cases presenting factor for the later development of type 1 diabetes, possibly between August and November) (McLarty, Yusafai, and Swai because, in neonates, bovine albumin can raise antibodies 1989). It would therefore seem likely that potential viral trig- that mimic islet cell antibodies and attack pancreatic beta gers operate also in the rest of Africa. cells. This is, of course, speculative, and much still remains unknown about the pathogenesis and epidemiology of type 1 diabetes in Africa. PREVALENCE OF TYPE 2 DIABETES Before the 1990s, diabetes was considered a rare medical Genetic Factors condition in Africa. Epidemiological studies carried out in More than 90 percent of type 1 diabetes subjects in Sub- that decade, however, provided evidence of a trend toward Saharan Africa, as in the rest of the world, have one or increased incidence and prevalence of type 2 diabetes in both human leukocyte antigens (HLA) DR3 and DR4. African populations (Sobngwi et al. 2001). Indeed, Africa is However, there appear to be specificities in the HLA sus- experiencing the most rapid demographic and epidemio- ceptibility found in certain African populations. Recent logical transition in world history (Mosley, Bobadilla, studies using allele-specific (oligonucleotide) probes from and Jamison 1993). It is characterized by a tremendous rise Zimbabwe, Senegal, and Cameroon show positive and in the burden of noncommunicable diseases (NCDs), 272 | Jean-Claude Mbanya and Kaushik Ramiaya underlined by the increasing life expectancy and lifestyle investigations in Sub-Saharan Africa. Such a need is dictated changes resulting from the reduction in infectious diseases by the prevalence of undiagnosed diabetes, which accounted and increased fertility, as well as Westernization. for 60 percent of those with diabetes in Cameroon (Mbanya Almost all the reports published between 1959 and 1985 et al. 1997), 70 percent in Ghana (Amoah, Owusu, and Adjei showed a prevalence of diabetes below 1.4 percent, except 2002), and over 80 percent in the recent study in Tanzania those from South Africa, where higher prevalence was (Aspray et al. 2000). It would therefore appear that in Sub- reported. Differences in diagnostic methods and criteria, Saharan Africa, for every diagnosed person with diabetes, however, made comparison between countries difficult. there are one to three undiagnosed cases. Since then, uniform diagnostic criteria has become more The International Diabetes Federation estimated that in available, allowing comparison across countries. 2003 the number of people age 20 to 79 years with diabetes Epidemiological studies carried out during this period show in Sub-Saharan Africa was over 7 million for a population of the rising prevalence of diabetes all over Africa (table 19.2). more than 295 million, giving a prevalence rate of 2.4 per- The prevalence of diabetes in Africa was approximately cent. About 65 percent of those affected with diabetes lived 3 million in 1994; but the region is due to experience a two- in the urban areas, whereas 35 percent lived in the rural to threefold increase by the year 2010 (Amos, McCarty, and communities. More than 46 percent (3.3 million) of the dia- Zimmet 1997). The highest prevalence is found in popula- betic population were 40 to 59 years old, whereas 28 percent tions of Indian origin, followed by black populations and and 26 percent, respectively, were 20 to 39 and 60 to 79 years Caucasians. Among the population of Indian origin in old (table 19.4). This has serious implications for the pro- South Africa and Tanzania, the prevalence is between 12 and ductivity of the region, since diabetes affects the active 13 percent (Ramaiya, Swai, McLarty, and Alberti 1991). The members of the community. The top five countries with prevalence in blacks follows a Westernization gradient, with the highest number of people affected by diabetes in Sub- that of rural Africa generally below 1 percent but that of Saharan Africa are Nigeria (about 1.2 million people), South urban Africa between 1 and 6 percent. In general the preva- Africa (841,000), the Democratic Republic of Congo lence of type 2 diabetes is low in both rural and urban com- (552,000), Ethiopia (550,000), and Tanzania (380,000) munities of West Africa except in urban Ghana, where a (table 19.4). high rate of 6.3 percent was recently reported (Amoah, The impact of type 2 diabetes is bound to continue if Owusu, and Adjei 2002). Moderate rates have been reported nothing is done to curb the rising prevalence of IGT, which from South Africa: 4.8 percent in a semi-urban community now varies between 2.2 percent and 16.2 percent. The esti- in the Orange Free State, 6.0 percent in an urban commu- mated prevalence rate of IGT in 2003 for Sub-Saharan nity of the Orange Free State, 5.5 percent in Durban (mostly Africa is 7.3 percent with a total population of affected indi- occupied by the Zulu tribe), and 8 percent in Cape Town viduals of over 21 million (table 19.5). Some 70 percent of (mostly occupied by the Xhosa tribe). Also, moderate rates these individuals are expected to develop type 2 diabetes have been reported in studies from Tanzania (table 19.1). unless something is done to reduce the risk factors associ- ated with the development of the disease. The projections of type 2 diabetes and IGT from 2003 to Estimates of the Prevalence of Type 2 Diabetes 2025 are shown in table 19.6. It is estimated that the burden and IGT in 2003 of diabetes and IGT will just about double in 2025 from There are marked discrepancies between the prevalence of their 2003 levels. The rate at which new cases of diabetes diabetes among different communities in Sub-Saharan are emerging poses an additional burden on countries Africa. The studies from Tanzania (Aspray et al. 2000; already stretched to the limit by common life-threatening McLarty et al. 1989), showing an urban-to-rural ratio of five infections, such as malaria, tuberculosis, and HIV and to one, and from Cameroon (Mbanya et al. 1997), with a acquired immune deficiency syndrome (AIDS). ratio of two to one, both confirm the urban-rural discrep- ancy in diabetes prevalence and suggest the consequent Risk Factors for Type 2 Diabetes likely increases because of urban migration. The data used for the extrapolation of current prevalence rates (table 19.2) There are marked differences between diabetic and nondia- in distant and probably dissimilar countries and popula- betic individuals in the prevalence of some risk factors for tions indicate the great need for more epidemiological diabetes and its complications, notably anthropometric Diabetes Mellitus | 273 Table 19.4 Prevalence Estimates of Diabetes Mellitus, by Country, 2003 Population DM Number of people with DM age 20­79 (thousands) age (20­79) prevalence Country (thousands) (%) Rural Urban Male Female Age 20­39 Age 40­59 Age 60­79 Total Angola 5,846 2.7 33.1 123.7 84.0 72.8 51.6 69.5 35.8 156.8 Benin 2,911 2.1 23.6 38.9 32.7 29.8 19.0 28.1 15.5 62.5 Botswana 716 3.6 3.2 22.3 8.8 16.7 4.1 13.9 7.5 25.5 Burkina Faso 4,969 2.7 90.4 44.8 67.8 67.4 40.1 55.2 40.0 135.3 Burundi 2,860 1.3 22.0 16.0 19.4 18.6 12.6 15.6 9.8 38.0 Cameroon 7,278 0.8 19.5 38.9 23.9 34.5 9.4 42.3 6.7 58.4 Cape Verde 228 2.3 1.1 4.2 2.2 3.1 2.0 1.7 1.7 5.3 Central African 1,780 2.3 16.3 25.0 21.1 20.2 10.7 17.6 13.0 41.3 Republic Chad 3,674 2.7 60.7 39.9 39.1 61.5 12.2 54.5 33.9 100.6 Comoros 355 2.5 1.9 7.0 4.8 4.1 3.2 3.9 1.8 8.9 Congo, Dem. Rep. of 22,436 2.5 136.6 415.4 294.7 257.3 182.3 237.4 132.3 552.0 Congo, Rep. of 1,403 2.6 7.6 28.3 18.4 17.6 10.5 15.2 10.3 35.9 Côte d'Ivoire 7,959 2.3 63.9 121.9 107.3 78.6 51.4 83.0 51.5 185.8 Djibouti 300 4.9 1.3 13.5 4.8 9.9 1.3 8.3 5.2 14.8 Equatorial Guinea 226 2.5 1.8 3.8 2.9 2.7 1.5 2.5 1.7 5.6 Eritrea 1,906 1.9 13.6 22.7 19.7 16.6 11.8 15.6 8.8 36.2 Ethiopia 29,562 1.9 214.6 335.8 299.4 250.9 176.6 234.9 138.8 550.4 Gabon 647 2.9 5.0 13.9 9.9 9.0 4.0 8.1 6.8 18.9 Gambia, The 703 2.2 7.4 8.0 8.3 7.0 4.1 7.2 4.0 15.4 Ghana 9,986 3.3 143.8 190.2 185.0 149.0 93.4 152.8 87.8 334.0 Guinea 3,855 2.0 37.6 41.2 42.8 36.0 23.3 35.5 20.1 78.9 Guinea-Bissau 588 2.0 7.0 4.8 6.3 5.5 3.1 5.2 3.5 11.8 Kenya 14,604 2.5 78.1 281.5 193.6 166.0 133.7 152.3 73.5 359.6 Lesotho 1,040 3.1 17.3 14.8 12.3 19.8 4.2 17.6 10.4 32.1 Liberia 1,573 2.0 10.5 21.3 17.0 14.8 11.6 11.6 8.6 31.8 Madagascar 7,782 2.5 47.3 144.6 104.3 87.5 63.2 85.5 43.2 191.9 Malawi 5,131 1.7 38.0 49.3 46.6 40.6 28.8 35.5 23.0 87.2 Mali 5,231 2.0 54.4 52.5 55.8 81.1 30.7 42.6 33.6 106.9 Mauritania 1,309 3.5 11.4 34.6 18.0 28.0 5.6 26.1 14.3 46.0 Mozambique 8,681 3.1 44.9 221.6 142.4 124.1 86.0 118.8 61.6 266.5 Namibia 831 3.1 10.1 15.4 9.5 16.0 3.9 13.6 8.0 25.5 Niger 4,728 3.1 36.4 110.3 57.7 89.0 20.8 83.9 41.9 146.7 Nigeria 54,248 2.2 439.3 779.4 655.4 563.3 354.5 528.9 335.3 1,218.7 Réuniona 474 13.1 10.0 51.9 29.1 32.8 11.2 29.7 21.0 61.9 Rwanda 3,645 1.1 28.3 13.1 22.7 18.7 15.2 14.7 11.4 41.4 São Tomé 107 2.8 1.0 1.9 1.6 1.4 0.6 1.3 1.0 2.9 and Principeb Senegal 4,607 2.3 34.9 68.9 54.7 49.0 31.3 46.8 25.6 103.7 Seychellesa,b 49 12.3 1.5 4.5 2.9 3.0 1.0 2.9 2.0 6.0 Sierra Leone 2,268 2.2 21.3 27.6 25.7 23.2 14.0 21.8 13.1 48.9 Somalia 4,086 2.3 24.5 67.5 49.6 42.4 32.2 40.5 19.3 92.0 South Africa 24,741 3.4 272.1 569.1 322.7 518.5 127.1 489.6 224.5 841.2 Swaziland 450 3.0 6.0 7.4 5.2 8.2 2.0 7.4 3.9 13.4 Tanzania 16,616 2.3 98.1 281.0 203.4 175.7 134.9 163.9 80.3 379.1 Togo 2,196 2.1 21.4 23.7 23.9 21.2 13.2 19.4 12.5 45.1 Uganda 10,018 1.5 71.0 83.9 84.9 70.0 56.7 60.7 37.5 154.9 Zambia 4,625 3.0 21.8 118.2 76.1 64.0 49.2 59.1 31.8 140.1 Zimbabwe 5,686 2.6 68.2 80.5 59.0 89.7 24.7 79.7 44.3 148.7 Total 295,065 2.4 2,380 4,692 3,580 3,491 1,985 3,265 1,821 7,072 Source: Adapted with permission from IDF 2003. Note: The total may not be the exact sum of the column due to the rounding. a. Réunion and the Seychelles were deemed to have the same ethnicity and distribution as Mauritius. b. Population number as described in the CIA World Factbook 2002, with age distribution adjustment to that of the world population in 2003. 274 | Jean-Claude Mbanya and Kaushik Ramiaya Table 19.5 Prevalence Estimates of IGT, by Country, 2003 Population IGT Number of people with IGT age 20­79 (thousands) age (20­79) prevalence Country (thousands) (%) Male Female Age 20­39 Age 40­59 Age 60­79 Total Angola 5,846 7.5 190.2 251.0 193.9 152.4 95.0 441.2 Benin 2,911 6.9 99.3 102.7 82.9 77.5 41.5 202.0 Botswana 716 7.0 31.4 18.9 15.3 12.3 22.7 50.3 Burkina Faso 4,969 7.0 156.8 189.3 144.6 113.2 88.3 346.1 Burundi 2,860 7.5 96.2 127.0 94.4 76.5 42.4 213.2 Cameroon 7,278 2.2 104.5 56.4 23.9 86.4 50.6 161.0 Cape Verde 228 6.8 6.2 9.2 7.2 4.2 4.0 15.4 Central African Republic 1,780 7.4 63.0 69.2 47.3 49.2 35.7 132.3 Chad 3,674 2.3 29.9 53.5 21.0 38.5 23.9 83.4 Comoros 355 7.3 11.3 14.7 12.4 8.7 4.9 26.0 Congo, Dem. Rep. of 22,436 7.6 729.1 966.7 746.0 572.8 377.0 1,695.8 Congo, Rep. of 1,403 7.2 48.7 51.7 39.5 36.5 24.4 100.4 Côte d'Ivoire 7,959 7.2 308.9 262.6 219.2 219.5 132.8 571.5 Djibouti 300 2.6 2.3 5.6 1.5 3.9 2.5 7.9 Equatorial Guinea 226 7.5 8.3 8.6 6.0 6.4 4.4 16.8 Eritrea 1,906 7.6 62.6 82.0 62.7 51.5 30.3 144.6 Ethiopia 29,562 7.6 978.6 1,270.7 966.5 796.1 486.7 2,249.3 Gabon 647 8.1 25.7 26.7 15.5 20.1 16.8 52.5 Gambia, The 703 7.4 26.0 25.8 18.9 21.5 11.5 51.9 Ghana 9,986 12.0 564.8 636.3 529.4 409.1 262.6 1,201.1 Guinea 3,855 7.0 137.9 133.9 109.0 104.9 57.8 271.7 Guinea-Bissau 588 7.4 21.4 22.0 15.9 16.7 10.8 43.4 Kenya 14,604 7.2 461.0 592.7 514.6 338.9 200.2 1,053.7 Lesotho 1,040 8.5 58.7 30.1 19.7 22.5 46.5 88.8 Liberia 1,573 6.5 52.0 50.6 49.8 30.4 22.4 102.7 Madagascar 7,782 7.5 256.3 331.0 257.2 207.1 122.9 587.3 Malawi 5,131 7.5 166.1 221.0 170.5 131.6 85.0 387.2 Mali 5,231 7.2 183.2 193.8 148.2 129.0 99.9 377.1 Mauritania 1,309 2.3 10.8 18.9 7.5 14.3 7.9 29.7 Mozambique 8,681 7.6 286.0 376.5 284.6 228.3 149.6 622.5 Namibia 831 8.0 43.3 22.9 17.0 15.4 33.7 66.1 Niger 4,728 6.7 160.8 157.1 140.8 117.7 59.4 318.0 Nigeria 54,248 7.1 1,947.9 1,900.5 1,527.5 1,432.8 888.1 3,848.4 Réuniona 474 16.2 29.2 47.6 29.2 31.0 16.6 76.8 Rwanda 3,645 7.2 114.7 149.5 128.5 83.3 52.4 264.2 São Tomé and Principeb 107 8.1 4.3 4.3 2.6 3.5 2.6 8.7 Senegal 4,607 7.0 160.5 162.2 131.1 124.7 55.9 322.7 Seychellesa,b 49 16.1 3.2 4.7 2.9 3.2 1.7 7.9 Sierra Leone 2,268 7.2 79.8 82.8 63.2 62.5 36.9 162.6 Somalia 4,086 7.4 129.9 171.1 139.6 104.5 56.9 301.0 South Africa 24,741 7.2 1,201.8 573.4 498.6 553.8 722.8 1,775.2 Swaziland 450 7.8 23.4 11.7 9.1 8.9 17.0 35.1 Tanzania 16,616 7.3 525.5 694.8 574.2 413.1 233.1 1,220.3 Togo 2,196 7.1 77.1 78.5 62.1 57.4 36.1 155.6 Uganda 10,018 7.3 319.3 407.9 351.2 233.9 142.1 727.2 Zambia 4,625 7.4 151.4 189.5 158.8 107.2 74.9 340.9 Zimbabwe 5,686 7.2 285.0 124.2 126.9 99.6 182.6 409.2 Total 295,065 7.3 10,426 10,984 8,789 7,434 5,186 21,410 Source: Adapted with permission from IDF 2003. Note: The totals may not be the exact sum of columns due to rounding. a. Réunion and the Seychelles were deemed to have the same ethnicity and distribution as Mauritius. b. Population number as described in the CIA World Fact Book 2002, with age distribution adjustment to that of the world population in 2003. Diabetes Mellitus | 275 Table 19.6 Projections of Diabetes and IGT from 2003 to 2025 countries it is in the age group 45 to 64, and in Sub-Saharan in the Age Group of 20 to 79 Years Africa it is in the age groups 20 to 44 and 45 to 64 years. Yet All diabetes and IGT 2003 2025 data from 12 other studies from Sub-Saharan Africa indicate two peak age ranges of 45 to 64 and older than 65 years (see Total population (millions) 666.6 1,107.4 table 19.1 for references). Adult population (millions) 295.1 541.1 Two studies in Sub-Saharan Africa have examined ethnic Diabetes prevalence (%) 2.4 2.8 differences in the prevalence of diabetes. A difference was Diabetes numbers (millions) 7.1 15.0 found between Indians, blacks, and Caucasians in South IGT prevalence (%) 7.3 7.3 Africa, where Indians had the highest predisposition and IGT numbers (millions) 21.4 39.4 were followed by blacks and Caucasians (Levitt et al. 1999; Source: Adapted with permission from IDF 2003. Omar et al. 1994). In the Tanzanian study, the indigenous African population had lower diabetes prevalence than variables, such as obesity. Although it is true that these data the migrant Asian group (1.1 percent as opposed to 9.1 to are from cross-sectional studies that have limitations in 7.1 percent) (McLarty et al. 1989; Ramaiya, Swai, McLarty, establishing causality, they at least support the hypothesis Bhopal, et al. 1991; Swai et al. 1990). that increasing prevalence of diabetes can be attributed The prevalence of diabetes appears to be substantially largely to changes in lifestyle resulting in reduced physical higher in African-origin populations living abroad than in activity and increased calorie intake and subsequent weight indigenous Africans. West Africans from Nigeria (Cooper gain. Such changes have important implications for the pro- et al. 1997) and central Africans from Cameroon (Mbanya vision of health care and for health education to promote et al. 1997) were compared with populations of West African behavioral change in order to control the emergence of dia- origin in the Caribbean (Cooper et al. 1997; Mbanya et al. betes in Sub-Saharan Africa. 1997), United Kingdom (Cooper et al. 1997; Mbanya et al. 1997), and the United States (Cooper et al. 1997). These Age and Ethnicity. Age and ethnicity are the two main studies suggest that environment determines diabetes nonmodifiable risk factors of diabetes in Africa. Glucose prevalence in these populations of similar genetic origin. intolerance in Sub-Saharan Africa, as in other regions of the world, increases with age in both men and women Urban-Rural Differences. Residence seems to be a major (figure 19.1); however, published studies lack uniformity determinant of diabetes in Sub-Saharan Africa, since urban on the age range in which the prevalence of diabetes is residents have 1.5- to 4.0 times higher prevalence of diabetes observed. According to King, Aubert, and Herman (1998), than their rural counterparts. This is attributable to lifestyle in most developed communities the peak of occurrence falls changes associated with urbanization and Westernization. in the age group of 65 years or older, whereas in developing Urban lifestyle in Africa is characterized by changes in dietary habits involving an increase in the consumption of refined sugars and saturated fat and a reduction in fiber Figure 19.1 Prevalence of Diabetes with Increasing Age intake (Mennen et al. 2000). Sobngwi and colleagues (2002) in Cameroon have recently reported an increase in fasting plasma glucose in those whose lives have been spent in an urban environ- 4.0 ment, suggesting that both lifetime exposure to and recent 3.5 migration to or current residence in an urban environment 3.0 are potential risk factors for obesity and diabetes mellitus. 2.5 The disease might represent the cumulative effects over 2.0 percent years of dietary changes, decrease in physical activity, and 1.5 psychological stress. 1.0 The population of Africa is predominantly rural, but the 0.5 1995­2000 urban growth rate was estimated at 4.3 percent 0 15­24 25­34 35­44 45­54 55 (compared with 0.5 percent in Europe). Thus, more than age 70 percent of the population of Africa will be urban resi- Source: Authors. dents by 2025 (UNFPA 2000). There will therefore be a 276 | Jean-Claude Mbanya and Kaushik Ramiaya tremendous increase in the prevalence of diabetes attribu- Figure 19.2 Mean Fasting Blood Glucose by Tertiles of table to rapid urbanization. In addition, life expectancy at Walking Energy Expenditure in Women: The Cameroon Study birth is rapidly increasing. For example, in Cameroon in 5.2 1960 it was about 35 years but in 1990 was raised to approx- 5.1 imately 55 years. An increase in diabetes prevalence simply 5.0 p for trend .006 p for trend .004 because of the change in the age structure of the population 4.9 liter 4.8 is therefore expected. However, the HIV pandemic may per 4.7 change these estimates and projections. 4.6 millimoles 4.5 Family History of Diabetes. A significant proportion of 4.4 the offspring of Cameroonians with type 2 diabetes have 4.3 4.2 either type 2 diabetes (4 percent) or IGT (8 percent) 30 30­49 50 age (Mbanya et al. 2000). A positive family history seems to be an independent risk factor for diabetes, but this was not 1st tertile 2nd tertile 3rd tertile the case in the Cape Town study (Levitt et al. 1993), in which Source: Adapted from Sobngwi, Gautier, and Mbanya 2003. family history was not an independent risk factor. Note: Third p for trend value was not shown in the original article. Measure of Adiposity. Several studies from Sub-Saharan sectional data from 1,417 women age 15 to 83 years in a Africa have confirmed the association between the preva- rural community and an urban community in Cameroon lence of diabetes and a surrogate of obesity, body mass index showed that in all age groups, fasting blood glucose levels (BMI). Reports from Mali (Fisch et al. 1987), Nigeria were inversely associated with energy expenditure from (Cooper et al. 1997) and Tanzania (McLarty et al. 1989) have walking (figure 19.2) (Sobngwi, Gautier, and Mbanya 2003). shown that the prevalence of diabetes increases with Rural dwellers' higher level of physical activity and related increasing BMI. BMI and obesity seem to be independent energy expenditure compared with urban subjects goes far risk factors for diabetes (Levitt et al. 1993). to explain why obesity was found to be at least four times higher in urban areas than rural (Aspray et al. 2000). Thus, Physical Activity. There seems to be a significant relation- lack of physical activity appears to be a significant risk fac- ship between physical inactivity and diabetes and obesity tor for diabetes in Sub-Saharan Africa. (Sobngwi et al. 2002). Physical activity is more common in rural than urban regions of Africa because rural populations rely on walking for transport and often have intense agri- COMPLICATIONS OF DIABETES cultural activities as their main occupation. In Sub-Saharan Africa, walking time and pace is drastically reduced (by fac- The escalating prevalence of type 1 and type 2 diabetes and tors of 2 to 4 for walking at a slow pace and 6 to more than their complications in Sub-Saharan Africa are a major drain 10 for walking at a brisk pace) in an urban community as on health resources in financially difficult circumstances, in compared with a rural community. The main difference in addition to having a considerable physical and social impact physical activity between the two types of community, how- on the individual and community. ever, is the use of walking in rural areas as a means of trans- portation. Acute Complications of Diabetes The reduction in physical activity associated with life in a city partly explains the excess prevalence of obesity in urban The three main metabolic complications of diabetes in areas. In a South African study, the prevalence of a sedentary Sub-Saharan Africa are diabetic ketoacidosis, hyperosmolar lifestyle in Cape Town in subjects age 30 years and over was nonketotic coma, and hypoglycemia. Diabetic ketoacidosis 39 percent for men and 44 percent for women (Omar et al. is a common diabetic emergency in developing countries 1993). Low physical activity was normal for 22 percent of and carries with it relatively high mortality, ranging from men and 52 percent of women in urban Tanzania, whereas 25 percent in Tanzania to 33 percent in Kenya. The major it was usual for only 10 percent of men and 15 percent of contributing factors to such high mortality are the chronic women living in rural areas (Edwards et al. 2000). Cross- lack of availability of insulin, delays in seeking medical Diabetes Mellitus | 277 assistance by newly diagnosed type 1 patients presenting in 15 percent of all cases (table 19.3). At diagnosis, 21 to 25 per- ketoacidosis, misdiagnosis of diabetes, and poor health care cent of type 2 patients and 9.5 percent of type 1 patients in general and diabetic care in particular (Rwiza, Swai, and have retinopathy. Ethnic differences in the prevalence of McLarty 1986). retinopathy have been observed in multiethnic communi- Hyperosmolar nonketotic coma is usually a complication ties. In South Africa, the highest prevalence of retinopathy is of type 2 diabetes and is less common and accounts for observed in Africans, rather than Indians or Caucasians about 10 percent of all hyperglycemic emergencies in devel- (whites), at diagnosis and after a similar duration of follow- oping countries (Zouvanis et al. 1997). Infection is the lead- up (Kalk et al. 1997). Although genetic predisposition may ing precipitating factor for both diabetic ketoacidosis and not be ruled out, lack of blood glucose and blood pressure hyperosmolar nonketotic coma, followed by first presenta- control because of difficult access to health care might tion of diabetes at a health institution and noncompliance account for most of these differences. with a medical regimen (Zouvanis et al. 1997). It carries a The prevalence of nephropathy varies between 32 and high mortality of up to 44 percent according to studies from 57 percent after a mean duration of diabetes of 5 to 10 years South Africa, which may be because the patients are usually and between 5 and 28 percent within the first year following elderly and have other major illnesses (Rolfe et al. 1995). the diagnosis of diabetes (table 19.3). Diabetic nephropathy Hypoglycemia is also a serious complication of treat- also occurs early in the course of diabetes, because between ment in patients with diabetes. Of a total of 51 episodes 32 and 57 percent of diabetic patients with a mean duration in 43 patients admitted at the Baragwanath Hospital, of diabetes between 5 and 10 years have microalbuminuria Johannesburg, South Africa, 14 cases (33 percent) were asso- (Kalk et al. 1997; Rahlenbeck and Gebre-Yohannes 1997; ciated with sulfonylurea treatment. The major cause precip- Sobngwi et al. 1999). The diagnosis of nephropathy may, itating the event was a missed meal (36 percent), although however, be faulty because of the presence of proteinuria alcohol (22 percent), gastrointestinal upset (20 percent), and due to renal infections and sickle-cell anemia. In Africa, dia- inappropriate treatment (18 percent) were also important betes mellitus accounts for a third of all patients who are contributory factors (Gill and Huddle 1993). No mortality admitted to dialysis units (Diallo et al. 1997), and renal was associated with hypoglycemia in this study. replacement is both expensive and not widely available. It appears, therefore, that diabetic end-stage renal failure is the first cause of hospital mortality in diabetic patients in Chronic Complications of Diabetes Africa. In South Africa, for example, 50 percent of all causes The seriousness of diabetes is largely a result of its associat- of mortality in type 1 diabetic patients may be due to renal ed complications, which can be serious, disabling, and even failure (Gill, Huddle, and Rolfe 1995). fatal. Prevalence studies on complications reported up to the The estimates of the prevalence of neuropathy vary wide- early 1990s gave widely variable figures. These have been ly, depending on the methodology used to assess them. reviewed in two studies and include figures ranging from Macrovascular complications of diabetes are considered 9 to 16 percent for cataract, 7 to 52 percent for retinopathy, rare in Africa despite a high prevalence of hypertension. 6 to 47 percent for neuropathy, 6 to 30 percent for Lower-extremity amputation varies from 1.5 to 7 percent, nephropathy, and 1 to 5 percent for macroangiopathy and about 12 percent of all hospitalized diabetic patients (Mbanya and Sobngwi 2003; Rolfe 1997). The variations are have foot ulceration. A high proportion of patients have due to diagnostic criteria problems, local and geographical lower-limb arterial disease that contributes to the develop- factors, type of diabetes, and variation in duration of dia- ment of diabetic foot lesions. It is common to see patients betes. Since 1995, however, many more vigorous and well- with diabetic foot ulcers as the presenting complaint of conducted studies have taken place, giving a much clearer diabetes. Data from Tanzania have shown that the vast picture of complication prevalence; these are summarized in majority (over 80 percent) of ulcers are neuropathic in table 19.3. It can be seen that there is generally less wide a origin and not associated with peripheral vascular disease range between these studies and also that the figures them- (Abbas, Lutale, and Morback 2000). Audits of diabetes care selves are substantial. carried out in Cape Town, South Africa; Dar es Salaam, The prevalence of diabetic retinopathy varies from 13 to Tanzania; and Yaoundé, Cameroon, have demonstrated 55 percent, depending on the duration of diabetes and poor glycemic control and inadequate foot care as risk glycemic control, with severe retinopathy representing factors for diabetic foot. Fewer than 22 percent of patients 278 | Jean-Claude Mbanya and Kaushik Ramiaya had their feet examined during a year of attendance at Swai, Lutale, and McLarty 1990). Eighty-two percent of primary health care clinics in these three cities, even though those not on insulin survived five years, but only 60 percent in Cape Town, 37 percent demonstrated either peripheral of the group on insulin treatment survived that long. Once neuropathy or peripheral vascular disease (Abbas, Lutale, again, the causes of death were predominantly metabolic and Morback 2000; Boulton 1990). Limited patient knowl- and infective. The authors concluded that in Africa "diabetes edge of proper foot care, practices relating to foot care, and was a serious disease with a poor prognosis." One reviewer cultural beliefs, including the association of diabetes, leg also observed that the Tanzanian study indicated that ulcers, and lower extremity amputation with bewitchment, five years from diagnosis, 40 percent of those on insulin are also common problems encountered in Sub-Saharan would die, whereas in Europe 40 percent of similar patients Africa countries (Abbas, Lutale, and Morback 2000; Boulton would survive more than 40 years (Deckert, Poulsen, and 1990). Larsen 1978; Gill 1997). Data on cerebrovascular disease are scarce because of the There is some evidence, however, that at least in some mortality associated with this complication, the low propor- parts of Africa the prognosis of diabetes is improving. tion of patients seen in hospitals, and the lack of death cer- Figures reported from Ethiopia (Lester 1991, 1996), for tificates or proper records of the cause of death. Recent example, are considerably better than the Zimbabwean results from the general population of Tanzania, where a (1980) and Tanzanian (1990) data. Interestingly, although morbidity and mortality surveillance system has been set metabolic emergencies were still the major cause of death, up, show that stroke mortality was three to six times that of the mortality from renal failure was substantial, presumably England and Wales and that 4.4 percent of type 2 diabetic from diabetic nephropathy and large vessel disease. A cohort patients presented with stroke at the diagnosis of diabetes of type 1 diabetic patients who were followed in Soweto, (Walker et al. 2000). Coronary heart disease may affect 5 to South Africa, has also shown relatively prolonged survival 8 percent of type 2 diabetic patients and cardiomyopathy up (Gill, Huddle, and Rolfe 1995). At follow-up after 10 years, to 50 percent of all patients. Whereas microvascular compli- with a mean diabetes duration of 14 years, only 16 percent cations of diabetes are highly preventable and occur early had died. This figure was still in excess of Western rates, during the course of the disease, macrovascular disease is although almost all these deaths were due to nephropathy, a rare. Late diagnosis of diabetes, poor metabolic control, and complication mostly untreatable in Africa even now. nonstandardized diagnostic procedures rather than genetic Obviously, many factors affect mortality patterns among predisposition may account for this difference with other diabetic patients in different parts of Africa. These include populations around the world. provision of medical care and supply of insulin and other treatment modalities, as well as a variety of social, cultural, and ethnic factors. As seen earlier, the gradual lengthening MORTALITY ASSOCIATED WITH DIABETES of the duration of diabetes in itself contributes to changing patterns of mortality, to which diabetic nephropathy and There have been relatively few structured mortality studies macroangiopathy are rapidly playing a larger part in many from Africa, making quantification of outcome difficult. areas. Large vessel disease is likely to accelerate in preva- However, a major study from Zimbabwe in 1980 (Castle and lence also because of Western influences, such as smoking, Wicks 1980) recorded follow-up of 107 newly diagnosed obesity, reduced exercise, and high-fat diets. diabetic patients (both type 1 and type 2). In-patient mor- tality was 8 percent, and the survivors had a mortality rate of 41 percent within six years of follow-up. Most deaths COST OF DIABETES were due to infection, hyperglycemic emergencies (ketoaci- dosis and nonketotic coma), or hypoglycemia. Particular Studies on the economics of diabetes care in Sub-Saharan risk factors for an adverse outcome were male gender, alco- Africa are limited. A Medline search of such studies over the hol abuse, and insulin treatment. past 20 years yielded only the Tanzanian study (Chale et al. Another outcome study was reported from Tanzania in 1992). In Tanzania about US$4 million would have been 1990. A cohort of 1,250 newly diagnosed patients was fol- required to take care of all patients with diabetes in 1989/90, lowed from 1981 to 1987, and actuarial five-year survival which translates to US$138 per patient per year. This sum rates were calculated (McLarty, Kinabo, and Swai 1990; is equivalent to 8.1 percent of the total budgeted health Diabetes Mellitus | 279 expenditure for that financial year and well above the allo- STRATEGIES FOR CONTROL cated per capita health expenditure in Tanzania of US$2 for the year 1989/90 (Chale et al. 1992). In Cameroon the aver- During the last 10 years, health care spending in the devel- age direct medical cost of treating a patient with diabetes in oping countries has remained low. Of the 40 heavily 2001 was US$489, of which 56 percent was spent on hospi- indebted poor countries (HIPC) defined by the World Bank, tal admissions, 33.5 percent on antidiabetic drugs, 5.5 per- 33 of them are in Sub-Saharan Africa. The average per capita cent on laboratory tests, and 4.5 percent on consultation income of HIPCs, is US$310 per year. The health care fees. The direct medical costs for treating all diabetic spending is approximately US$8 per person per year, and patients in Cameroon represented about 3.5 percent of the pharmaceutical spending is approximately US$2 to $3 per national budget for the year 2001/2002 (Nkegoum 2002). person per year (WHO 2003). Estimates of diabetes care management in Malawi, based on The economic cost of diabetes and its complications is international prices for essential drugs and Malawi hospital unaffordable by most Sub-Saharan Africans. Their incomes cost data, suggest that a type 1 diabetic patient spends about are insufficient to purchase insulin, oral hypoglycemic US$100 per year for the purchase of insulin, and a type 2 agents, and other supplies for management of diabetes. The patient spends US$25 annually on oral hypoglycemic agents limited resources of the countries in Sub-Saharan Africa are (Vaughan, Gilson, and Mills 1989). divided between fighting poverty, implementing education The average age at onset of diabetes in Tanzania was strategies, providing housing and appropriate sanitation, 44 years and the average age at death was 46 years; popula- and dealing with the socioeconomic and health burden tion life expectancy was 53 years. The calculated number of of fighting the increasing incidence and prevalence of healthy life days (HLDs) lost because of diabetes was HIV/AIDS. Diabetes poses an additional burden on the lim- 4,100 days per patient, of which 69 percent was because of ited health care delivery system. premature mortality. This calculation was based on an aver- The problems encountered in the management of dia- age case-fatality rate after five years of 29 percent and a betes in Sub-Saharan Africa include diagnosis; medical care; severe chronic disablement rate of 14 percent. The estimated insulin and other drug supplies; monitoring; infections HLDs lost per capita because of diabetes were 820 person- associated with diabetes, especially the diabetic foot; dietary days per 1,000 people per year (Chale et al. 1992). In Ghana advice; diabetes education; and the low priority placed on the average age at onset of diabetes was 40 years, with noncommunicable diseases. 50 percent case fatality after 15 years and an average age at The training of health care providers and organizations is death of 55 years, with 30 percent disablement before death. not focused on effective and efficient treatment of people The total days lost were calculated as 217 per 1,000 people with diabetes. With modernization, economic well-being, per year, of which 52 percent were due to premature death and a Westernized lifestyle, the burden of diabetes and its (Vaughan, Gilson, and Mills 1989). complications also increases significantly. The resource- Table 19.7 reports the calculated estimates of the costs of limited countries are unable to provide even minimum care diabetes care in Sub-Saharan Africa for persons age 20 to in some instances, let alone secondary and tertiary care. 79 years (IDF 2003). The table uses population estimates Over the last 10 years several assessments of health care and diabetes prevalence estimates reported in table 19.4 for services for diabetes have been done, particularly in South those age 20 to 79 years for 2003. The total health care budg- Africa (Whiting, Hayes, and Unwin 2003). The findings et for 20- to 79-year-olds can be derived by multiplying the from these studies have shown the following: population figures for that age group by the per capita health expenditures. Calculations are then presented for val- · Patients' attendance is poor. ues of R of 2 and 3. R is the ratio of the cost of care for · Consultation times are short, resulting in little or no time people with diabetes compared with the cost of care of peo- for patient education. ple without diabetes. The data suggest that, at least for coun- · Staffing levels are inadequate, and staffs' knowledge is tries with high or moderate incomes, the value of R lies used inappropriately. between 2 and 3 (IDF 2003). The top five countries with · Staff are poorly or inadequately trained, or both, and the highest costs of diabetes care in Sub-Saharan Africa there exist hardly any continuous education programs. are South Africa, Kenya, Zimbabwe, Nigeria, and Ghana · Monitoring and evaluation of complications of diabetes (table 19.7). are lacking. 280 | Jean-Claude Mbanya and Kaushik Ramiaya Table 19.7 Calculated Estimates of the Costs of Diabetes Care, by Country Cost of diabetes care per year given Overall per capita values for Rb (thousands of international $a) health expenditure Country (international $a) R 2 R 3 Angola 52 7,940.6 15,477.0 Benin 27 1,652.0 3,236.0 Botswana 358 8,815.1 17,044.0 Burkina Faso 37 4,873.4 9,495.1 Burundi 16 600.0 1,184.5 Cameroon 55 3,186.4 6,322.5 Cape Verde 92 476.5 932.0 Central African Republic 37 1,493.4 2,920.6 Chad 19 1,860.5 3,624.3 Comoros 35 303.9 593.3 Congo, Dem. Rep. of 21 11,313.6 22,096.7 Congo, Rep. of 25 875.1 1,737.6 Côte d'Ivoire 45 8,170.3 15,976.1 Djibouti 63 888.6 1,697.4 Equatorial Guinea 103 562.8 1,099.1 Eritrea 25 888.1 1,743.8 Ethiopia 17 9,182.5 18,035.5 Gabon 171 3,140.2 6,107.1 Gambia, The 46 693.2 1,357.4 Ghana 51 16,482.7 31,932.0 Guinea 56 4,329.8 8,489.3 Guinea-Bissau 28 323.9 653.3 Kenya 115 40,360.2 78,826.2 Lesotho 100 3,113.9 6,046.6 Liberia 3 93.5 183.4 Madagascar 33 6,180.3 12,070.1 Malawi 38 3,258.2 6,409.4 Mali 32 3,352.3 6,573.0 Mauritania 52 2,310.8 4,469.9 Mozambique 30 7,756.9 15,065.0 Namibia 366 9,055.1 17,586.6 Niger 22 3,130.3 6,077.7 Nigeria 20 23,838.5 46,651.9 Réunion -- Rwanda 40 1,637.4 3,238.4 São Tomé and Principe 23 64.9 126.5 Senegal 56 5,679.4 11,114.1 Seychelles 758 4,049.6 7,299.4 Sierra Leone 28 1,340.3 2,625.2 Somalia 7 629.8 1,232.5 South Africa 663 539,377.0 1,044,412.1 Swaziland 210 2,732.7 5,311.9 Tanzania -- Togo 36 5,491.5 10,818.3 Uganda 36 5,491.5 10,818.3 Zambia 49 6,663.1 12,945.5 Zimbabwe 171 24,779.6 48,327.5 Total 784,539 1,522,237 Source: Data on per capita health expenditure are from WHO 2003. Adapted with permission from IDF 2003. Note: -- not available. a. The international dollar is a common currency unit that takes into account differences in the relative purchasing power of various currencies. Figures expressed in international dollars are calculated using purchasing power parities (PPP), which are rates of currency conversion constructed to account for differences in price level between countries. b. R ratio of cost of care for people with diabetes and without diabetes. Diabetes Mellitus | 281 · The control of blood glucose and blood pressure is poor Figure 19.3 Cost of Different Types of Insulin in Relation to the Gross National Product and inadequate. · Referral systems are almost nonexistent. 14 · Education of people with diabetes is lacking. 12 · Overall organization of the clinics is not satisfactory. $) 10 · Record keeping is poor. (US 8 · Even if treatment guidelines are available, they are hardly insulin used and are not up to date. 6 of · Health care systems in Sub-Saharan Africa vary widely. cost 4 2 A structured, organized diabetes health care system is 0 lacking. Many people with diabetes are managed by tradi- Africa EMME Europe North SACA Southeast Western America Asia Pacific tional health care providers and general practitioners who region are inadequately integrated into the primary care system. Human insulin Pork insulin Beef insulin Other problems with or barriers to the quality of delivery and affordable care include the following: Source: Adapted from IDF 2003. Note: EMME Eastern Mediterranean and Middle East; SACA South and Central America. · inadequate infrastructure · irregular supply of medicines · unaffordable insulin, oral hypoglycemic agents, and anti- Box 19.1 Organization of Diabetes Care hypertensives Problem Solution · disproportionate distribution of health care facilities 1. Lack of resources Diabetes associations · lack of information and clear roles for members of dia- Financial Twinning of diabetes betes health care teams Material associations · lack of appropriate and locally adapted diabetes edu- Support groups cation programs for people with diabetes and diabetes 2. Prevalence Research, studies health care professionals Magnitude of problem · lack of government support or subsidy, resulting in unaf- Disease complications fordable costs. 3. Lack of recognition of Communications interdependence of Transport Cost of medication, especially the high cost of insulin, is various levels of health Trained personnel a major handicap to proper diabetes care in Sub-Saharan care delivery Africa. Indeed, in a recent International Diabetes Federation Referrals Drugs, equipment survey (IDF 2003), it was observed that 80 percent of the supply people with diabetes were unable to obtain insulin and 4. Personnel Trained manpower insulin syringes because they could not afford them. The Motivation cost of insulin preparations was higher in Sub-Saharan Continuous Africa than elsewhere (figure 19.3). Insulin and insulin education syringes were accessible to only 11 percent of all people with Interpersonal diabetes in Africa. In addition, only 25 percent of people communication with diabetes monitored their blood glucose. Self- Migration monitoring of blood glucose was rarely used, mainly 5. Alternative health Confidence because of the cost of testing supplies in 90 percent and the care provision Cooperation unavailability of testing supplies in 70 percent of the coun- Scientific research tries in Africa (IDF 2003). 6. Compliance Education It is clear that the organization of diabetes care in Sub- Transport Saharan Africa has limitations at several levels of health Referrals delivery (box 19.1). There are solutions for each level of Source: Authors. limitations and problems. Implementation strategies for 282 | Jean-Claude Mbanya and Kaushik Ramiaya effecting change will vary from country to country based on focused education and mass-media campaigns, are highly several factors. Care for people with diabetes needs to begin cost-effective (Swai et al. 1990). Prevention strategies in at the primary health care level, followed by secondary and Sub-Saharan Africa have their own limitations. Lack of tertiary health care. In Sub-Saharan Africa, the rural-to- awareness by the population of and facilities for detection urban migration results in a significant increase in the geo- and monitoring contributes to the high prevalence of dia- graphical spread of people with diabetes. The health care for betic complications, and poorly skilled or inadequate health people with diabetes tends to be concentrated in large urban care staff, delay in seeking medical attention, and lack of hospitals, too far from most of the people who need to use access to affordable drugs contribute to the high rate of the services. Outreach programs in which health care diabetes-related mortality. Unless these factors are taken workers move out into the community are feasible, afford- into account when planning effective preventive strategies, able, and achievable and will benefit the rural population. the objectives will not be attainable. That will probably be the long-term solution to the prob- The United Kingdom Prospective Diabetes Study lems of the management of diabetes. (UKPDS 1998a) and the Diabetes Control and Complication Trial (DCCT 1993) have shown that intensive control of glucose results in a 25 to 70 percent reduction in the number DIABETES HEALTH CARE and severity of microvascular complications in people with diabetes. The UKPDS also demonstrated a 12 percent reduc- More than 50 years ago people with diabetes were mostly tion in mortality related to type 2 diabetes. In the UKPDS, treated in hospitals by specialists. With limited resources control of high blood pressure reduced the risk of microvas- and shrinking health budgets, together with a sharp rise in cular complications by 37 percent and death from type 2 the prevalence of type 2 diabetes, specialist care in a hospital diabetes­related disease by 32 percent (UKPDS 1998b), is not possible for everyone. An increasing number of pri- better reductions than those from tight blood glucose con- mary and community health care professionals are respon- trol (UKPDS 1998a), although the combination of blood sible for managing people with diabetes. The ongoing health pressure and blood glucose control was the most effective. sector reforms in most of the countries of Sub-Saharan Health beliefs are still deeply enshrined in the healing Africa have promoted more responsive and appropriate cultures of people in Sub-Saharan Africa, thereby predis- planning through decentralized and demand-driven health posing most patients to alternate between modern and tra- care, in which district managers decide how to divide their ditional clinics. There are places in Sub-Saharan Africa budgets between prevention and control of different health where chlorpropamide and tolbutamides are still drugs of problems (WHO 2000). choice (and the only ones available) together with alpha In order to plan for proper delivery of diabetes health methyldopa (for blood pressure control) in people affected care, Sub-Saharan Africa countries need epidemiological with diabetes. The newer classes of drugs--sulfonylurea and health services information. There is also a need to group, glinides--are unaffordable for the majority of the know the estimates of the prevalence of diabetes, its risk fac- population. tors, and its complications. Finally, adequate knowledge of One of the major challenges facing insulin-treated the overall burden of diabetes in high-risk populations and patients in Sub-Saharan Africa is the lack of a constant sup- countries is a prerequisite for effective diabetes health care ply of insulin at affordable cost (Yudkin 2000). The supply delivery (King et al. 1998). of insulin in Sub-Saharan Africa is erratic, even at large hos- pitals, and the prospects for people with type 1 diabetes are poor (Amoah et al. 1998; Dagogo-Jack 1995). The exact bur- Prevention den of poor insulin access in developing countries is still Sufficient evidence exists from countries outside Africa that unknown, because no good scientific study has been carried weight loss, diet, and exercise can prevent or delay diabetes out in these countries. However, 16 percent of the world's in people with IGT (Pan et al. 1997; Tuomilehto et al. 2001; population in developed countries with about 35 percent of Vijan et al. 1997) and that physical activity may exert an all diabetic patients use over 40 percent of the world's total independent effect on the prevention and control of dia- insulin each year (Jervell 1996; King 1998). Moreover, a betes (Wojtaszewski et al. 2000). small percentage of type 2 diabetes patients in developing The strategies for primary prevention, mostly involving countries require insulin when they become severely wasted support for behavior change through different forms of and hyperglycemic. Therefore, insulin is underutilized in Diabetes Mellitus | 283 Sub-Saharan Africa. In the second edition of Diabetes Atlas, The Role of the Patient the IDF's Task Force on Insulin Survey reports that no coun- Patients must be empowered and motivated to join associa- try in Africa had 100 percent accessibility to insulin. In fact tions. They have to be informed about their rights. Together two countries in Africa had the lowest accessibility in the with the community, they would be able to remove mis- world: the Democratic Republic of Congo, where people conceptions, mistrust, and the stigma of diabetes in Sub- with type 1 diabetes had access to insulin for less than Saharan Africa. The standard and quality of care being 25 percent of the time, and Zambia, where those with type 2 provided for people with diabetes has many limitations, diabetes had access to insulin only 26 to 49 percent of the which have to be overcome by a multisectoral approach. time. The high cost of insulin appears to be the most impor- tant cause of lack of access to insulin in people with type 1 diabetes in most countries of Africa (IDF 2003). There is CONCLUSION therefore an urgent need for the initiation of international programs to alleviate the plight of insulin-treated patients in Epidemiological data on diabetes mellitus in Sub-Saharan Africa. These programs may include (a) a selection of type Africa are still limited. However, the prevalence and inci- of drug through the essential drug list; (b) improving dence of both type 1 and type 2 diabetes are increasing with affordability of the price charged by applying such measures the persistent rural-to-urban migration. Clearly, knowledge as national price information, patent status, availability of of the disease has increased since the 1990s; nevertheless, generics, equity pricing schemes, review of general taxes and adoption of a Western lifestyle has greatly enhanced its margins, and as a last resort, parallel import and compulsive development. The evidence shows that type 2 diabetes is a licensing; and (c) sustainable financing of medical supplies major cause of morbidity and mortality on the continent through general tax levies, insurance schemes, copayment and that it is costly to manage diabetes and its complica- or full payment by the patient, loans, and donations (IDF tions. Given that the region still has a double and sometimes 1998). One of the solutions to the problem being discussed triple disease burden and that little priority is given to non- is donation of insulin combined with a mechanism to sup- communicable diseases like diabetes, and in the absence of a port logistics, education, and monitoring. The limiting fac- health care system adapted to this new reality and able to use tor with this scheme is long-term sustainability. costly therapeutic interventions, well-planned cost-effective Very few countries in Sub-Saharan Africa can afford to methods of prevention and treatment and refined tools to screen and treat the complications of diabetes (nephropa- assess health services and monitor progress are therefore thy, retinopathy, neuropathy, peripheral vascular disease) required. (Dagogo-Jack 1995). ACE inhibitors are a cost-effective way to reduce mortality and end-stage renal failure in people with type 1 diabetes (and type 2 diabetes) with microalbu- minuria, but they are of limited use because most people REFERENCES cannot afford them (Hendry et al. 1997; Tooke, Thomas, Abbas, Z., J. Lutale, and S. Morback. 2000. "Outcome of Hospital and Viberti 2000). The potential for intervention and Admissions for Diabetic Foot Lesions at Muhimbili Medical Centre, prevention of diabetic foot lesions is very high as a cost- Dar es Salaam, Tanzania." Diabetologia 43 (Suppl. 1): A246. effective strategy. Afoke, A. O., N. M. Ejeh, E. N. Nwonu, C. O. Okafor, N. J. Udeh, and J. Ludvigsson. 1992. "Prevalence and Clinical Picture of IDDM in Nigerian Ibo School Children." Diabetes Care 15: 1310­12. Ahren, B., and C. B. Corrigan. 1984. "Prevalence of Diabetes Mellitus in The Role of Diabetes Associations North-Western Tanzania." Diabetologia 26: 333­36. Amoah, A. G., S. K. Owusu, and S. Adjei. 2002. "Diabetes in Ghana: A The role of Diabetes Associations cannot be overempha- Community Based Prevalence Study in Greater Accra." Diabetes sized. National Diabetes Associations should urge the gov- Research and Clinical Practice 56: 197­205. ernment and nongovernmental organizations to ensure Amoah, A. G., S. K. Owusu, J. T. Saunders, W. L. Fang, H. A. Asare, J. G. Pastors, C. Sanborn, E. J. Barrett, M. K. Woode, and K. Osei. 1998. consistent and readily available insulin and other antidia- "Facilities and Resources for Diabetes Care at Regional Health betes drugs and supplies at a subsidized cost. The national Facilities in Southern Ghana." Diabetes Research and Clinical Practice associations can also play a significant role in imparting 42: 123­30. Amos, A. F., D. J. McCarty, and P. Zimmet. 1997. "The Rising Global education at different levels of health care and to the com- Burden of Diabetes and Its Complications: Estimates and Projections munity at large. to the Year 2010." Diabetic Medicine 14: S1­85. 284 | Jean-Claude Mbanya and Kaushik Ramiaya Aspray, T. J., F. Mugusi, S. Rashid, D. Whiting, R. Edwards, K. G. Alberti, Alleles in Black IDDM Controls from Zimbabwe." Tissue Antigens 40: and N. C. Unwin. 2000. "Rural and Urban Differences in Diabetes 145­49. Prevalence in Tanzania: The Role of Obesity, Physical Inactivity and Gill, G. V. 1997. "Outcome of Diabetes in Africa." In Diabetes in Africa, ed. Urban Living." Transactions of the Royal Society of Tropical Medicine G. V. Gill, J.-C. Mbanya, and K. G. M. M. Alberti, 65­71. Cambridge: and Hygiene 94: 637­44. FSG Communications. Boulton, A. 1990. Practical Diabetes Digest 1: 35­37. Gill, G. V., and K. R. Huddle. 1993. "Hypoglycaemic Admissions among Castle, W. N., and A. C. B. Wicks. 1980."Follow-Up of 93 Newly Diagnosed Diabetic Patients in Soweto, South Africa."Diabetic Medicine 10: 181­83. African Diabetics for 6 Years." Diabetologia 18: 121­23. Gill, G. V., K. R. Huddle, and M. Rolfe. 1995. "Mortality and Outcome of Chale, S. S., A. B. M. Swai, P. G. M. Mujinja, and D. G. McLarty. 1992."Must Insulin-Dependent Diabetes in Soweto, South Africa." Diabetic Diabetes Be a Fatal Disease in Africa? Study of Costs of Treatment." Medicine 12: 546­50. British Medical Journal 304: 1215­18. Hendry, B. M., G. C. Viberti, S. Hummel, A. Bagust, and J. Piercy. 1997. Chauffert, M., A. Cisse, D. Cevenne, B. Parfait, S. Michel, and F. Trivin. "Modelling and Costing the Consequences of Using an ACE Inhibitor 1995. "HLA-DR Beta 1 Typing and Non-Asp 57 Alleles in the to Slow the Progression of Renal Failure in Type I Diabetic Patients." Aborigine Population of Senegal." Diabetes Care 18: 677­90. Quarterly Journal of Medicine 90 (4): 277­82. CIA (Central Intelligence Agency). 2002. World Factbook 2002. Hunter, J. M., B. T. Sparks, J. Mufunda, C. T. Musabayane, H. V. Sparks, Washington, DC: CIA. http://www.cia.gov/publications/factbook/ and K. Mahomed. 2000. "Economic Development and Women's Blood index.html. Pressure: Field Evidence from Rural Mashonaland, Zimbabwe." Social Cook, A. R. 1901. "Notes on the Diseases Met with in Uganda, Central Science and Medicine 50: 773­95. Africa." Journal of Tropical Medicine 4: 175­78. IDF (International Diabetes Federation). 1998. Access to Insulin: A Report Cooper, R., C. Rotimi, J. Kaufman, E. Owoaje, H. Fraser, T. Forester, on the IDF Insulin Task Force on Insulin, 1994­1997. Brussels: IDF. R. Wilks, L. K. Riste, and J. K. Cruikshank. 1997."Prevalence of NIDDM ------. 2003. Diabetes Atlas. 2nd ed. Brussels: IDF. among Populations of the African Diaspora." Diabetes Care 20: 343­48. Jervell, J. 1996. "Variations in the Utilization and Cost of Insulin." Dagogo-Jack, S. 1995. "DCCT Results and Diabetes Care in Developing International Diabetes Federation Bulletin 41: 1­2. Countries." Diabetes Care 18: 416­17. Kalk, W. J., K. R. L. Huddle, and F. J. Raal. 1993. "The Age of Onset of DCCT Research Group. 1993. "The Effect of Intensive Treatment of Insulin-Dependent Diabetes Mellitus in Africans in South Africa." Diabetes on the Development and Progression of Long-Term Postgraduate Medical Journal 69: 552­56. Complications of Insulin Dependent Diabetes Mellitus." New England Journal of Medicine 329: 977­86. Kalk, W. J., J. Joannou, S. Ntsepo, I. Mahomed, P. Mahanlal, and P. J. Becker. 1997. "Ethnic Differences in the Clinical and Laboratory Deckert, T., J. E. Poulsen, and M. Larsen. 1978. "Prognosis of Diabetics Associations with Retinopathy in Adult Onset Diabetes: Studies in with Diabetes Onset before the Age of Thirty-One. I. Survival, Causes Patients of African, European and Indian Origins." Journal of Internal of Death, and Complications." Diabetologia 14: 363­70. Medicine 241: 31­37. Diallo, A. D., D. Nochy, E. Niamkey, and B. Yao Beda. 1997. "Etiologic Kaufman, J. S., E. E. Owoaje, C. N. Rotimi, and R. S. Cooper. 1999. "Blood Aspects of Nephrotic Syndrome in Black African Adults in a Hospital Pressure Change in Africa: Case Study from Nigeria." Human Biology Setting in Abidjan." Bulletin Société Pathologie Exotique 90: 342­45. 71 (4): 641­57. Dowse, G. K., H. Gareeboo, P. Z. Zimmet, K. G. Alberti, J. Tuomilehto, King, H. 1998. "Insulin: Availability, Affordability, and Harmonization." D. Fareed, L. G. Brissonnette, and C. F. Finch. 1990. "High Prevalence WHO Drug Information 12(4): 230­34. of NIDDM and Impaired Glucose Tolerance in Indian, Creole, and Chinese Mauritians." Mauritius Noncommunicable Disease Study King, H., R. E. Aubert, and W. H. Herman. 1998. "Global Burden of Group. Diabetes 39: 390­96. Diabetes. Prevalence, Numerical Estimates and Projections." Diabetes Care 21: 1414­31. Drabo, P. P. Y., J. Kabore, and A. Lengani. 1996. "Complications of Diabetes Mellitus at the Hospital Center of Ouagadougou." Bulletin Société Kitange, H. M., H. Machibya, J. Black, D. M. Mtasiwa, G. Masuki, D. Pathologie Exotique 89: 191­95. Whiting, N. Unwin et al. 1996. "The Outlook for Survivors of Childhood in Sub-Saharan Africa: Adult Mortality in Tanzania." British Ducorps, M., S. Baleynaud, H. Mayaudon, C. Castagne, and B. Bauduceau. Medical Journal 312: 216­20. 1996. "A Prevalence Survey of Diabetes in Mauritania." Diabetes Care 19: 761­63. Lester, F. T. 1984. "The Clinical Pattern of Diabetes in Ethiopians." Diabetes Care 7: 6­11. Edwards, R., N. Unwin, F. Mugusi, D. Whiting, S. Rashid, J. Kissima, T. J. Aspray, and K. G. Alberti. 2000. "Hypertension Prevalence and Care in ------. 1991. Clinical Status of Ethiopian Diabetic Patients after 20 Years an Urban and Rural Area of Tanzania." Journal of Hypertension 18 (2): of Diabetes." Diabetic Medicine 8: 272­76. 145­52. ------. 1993. "Clinical Features, Complications and Mortality in Type 2 Elamin, A., M. I. Omer, K. Zein, and T. Tuvemo. 1992. "Epidemiology of (Non-Insulin Dependent) Diabetic Patients in Addis Abeba, Ethiopia, Childhood Type I Diabetes in Sudan, 1987­1990." Diabetes Care 15: 1976­1990." Ethiopian Medical Journal 31 (2): 109­26. 1556­59. ------. 1996. "Mortality During Eight Years in a Cohort of Middle-Aged Elbagir, M. N., M. A. Eltom, E. M. Elmahadi, I. M. Kadam, and C. Berne. Ethiopian Diabetic Patients." International Diabetes Digest 7: 11­17. 1996. "A Population-Based Study of the Prevalence of Diabetes and Levitt, N. S., D. Bradshaw, M. F. Zwarenstein, A. A. Bawa, and S. Impaired Glucose Tolerance in Adults in Northern Sudan." Diabetes Maphumolo. 1997. "Audit of Public Sector Primary Diabetes Care Care 19: 1126­1128. in Cape Town, South Africa: High Prevalence of Complications in Fisch, A., E. Pichard, T. Prazuck, H. Leblanc, Y. Sidibe, and G. Brucker. Controlled Hyperglycaemia and Hypertension." Diabetic Medicine 14: 1987. "Prevalence and Risk Factors of Diabetes Mellitus in the Rural 1073­77. Region of Mali West Africa: A Practical Approach." Diabetologia 30: Levitt, N. S., J. M. Katzenellenbogen, D. Bradshaw, M. N. Hoffman, and 859­62. F. Bonnici. 1993 . "The Prevalence and Identification of Risk Factors Garcia-Pacheco, J. M., B. Herbut, S. Culbush, G. A. Hitman, W. Zhonglin, for NIDDM in Urban Africans in Cape Town, South Africa." Diabetes M. Magzoub, et al. 1992. "Distribution of HLA-DQA1 and DRBI Care 16: 601­7. Diabetes Mellitus | 285 Levitt, N. S., K. Steyn, E. V. Lambert, G. Reagan, C. J. Lombard, J. M. Fourie, in a Group of South African Blacks." South African Medical Journal 83: K. Rossouw, and M. Hoffman. 1999. "Modifiable Risk Factors for 641­43. Type 2 Diabetes Mellitus in a Peri-Urban Community in South Africa." Pan, X. R., G. W. Li, Y. H. Hu, J. X. Wang, W. Y. Yang, Z. X. An, Z. X. Hu, Diabetic Medicine 16: 946­50. et al. 1997. "Effects of Diet and Exercise in Preventing NIDDM in Mbanya, J. C., J. Ngogang, J. N. Salah, E. Minkoulou, and B. Balkau. People with Impaired Glucose Tolerance. The Da Qing IGT and 1997. "Prevalence of NIDDM and Impaired Glucose Tolerance in a Diabetes Study." Diabetes Care 20: 537­44. Rural and an Urban Population in Cameroon." Diabetologia 40: Panz, V. R., and B. I. Joffe. 1999. "Impact of HIV Infection and AIDS on 824­29. Prevalence of Type 2 Diabetes in South Africa in 2010." British Medical Mbanya, J. C., L. N. Pani, D. N. Mbanya, E. Sobngwi, and J. Ngogang. 2000. Journal 318: 1351. "Reduced Insulin Secretion in Offspring of African Type 2 Diabetic Pavan, L., E. Casiglia, P. Pauletto, S. L. Batista, G. Ginocchio, M. M. Y. Parents." Diabetes Care 23: 1761­65. Kwankam, et al. 1997. "Blood Pressure, Serum Cholesterol and Mbanya, J. C., and E. Sobngwi. 2003. "Diabetes Microvascular and Nutritional State in Tanzania and in the Amazon: Comparison with an Macrovascular Disease in Africa." Journal of Cardiovascular Risk 10: Italian Population." Journal of Hypertension 15: 1083­90. 97­102. Rahlenbeck, S. I., and A. Gebre-Yohannes. 1997. "Prevalence and McLarty, D. G., L. Kinabo, and A. B. M. Swai. 1990. "Diabetes in Tropical Epidemiology of Micro- and Macroalbuminuria in Ethiopian Diabetic Africa: A Prospective Study, 1981­7. II. Course and Prognosis." British Patients." Journal of Diabetes Complications 11: 343­49. Medical Journal 300: 1107­10. Ramaiya, K. L., A. B. M. Swai, D. G. McLarty, and K. G. M. M. Alberti. 1991. McLarty, D. G., A. B. M. Swai, H. M. Kitange, G. Masuki, B. L. Mtinangi, "Impaired Glucose Tolerance and Diabetes Mellitus in Hindu Indian P. M. Kilima, W. J. MaKene, L. M. Chuwa, and K. G. Alberti. 1989. Immigrants in Dar es Salaam." Diabetic Medicine 8: 738­44. "Prevalence of Diabetes and Impaired Glucose Tolerance in Rural Ramaiya, K. L., A. B. M. Swai, D. G. McLarty, R. S. Bhopal, and K. G. M. M. Tanzania." Lancet (8643): 871­75. Alberti. 1991. "Prevalences of Diabetes and Cardiovascular Disease McLarty, D. G., A. Yusafai, and A. B. M. Swai. 1989. "Seasonal Incidence of Risk Factors in Hindu Indian Subcommunities in Tanzania." British Diabetes Mellitus in Tropical Africa." Diabetic Medicine 6: 762­65. Medical Journal 303: 271­76. Mennen, L. I., J. C. Mbanya, J. Cade, B. Balkau, S. Sharma, S. Chungong, Rolfe, M. 1997. "Chronic Complications of Diabetes in Africa." In Diabetes and J. Kennedy 2000. "The Habitual Diet in Rural and Urban in Africa, ed. G. V. Gill, J.-C. Mbanya, and K. G. M. M. Alberti, 43­50. Cameroon." European Journal of Clinical Nutrition 24: 882­87. Cambridge: FSG Communications. Mollentze, W. F., A. J. Moore, A. F. Steyn, G. Joubert, K. Steyn, G. M. ------. 1988. "The Neurology of Diabetes Mellitus in Central Africa." Oosthuizen, and D. J. V. Weich. 1995. "Coronary Heart Disease Risk Diabetic Medicine 5: 399­401. Factors in a Rural and Urban Orange Free State Black Population." Rolfe, M., G. G. Ephraim, D. C. Lincoln, and K. R. L. Huddle. 1995. South African Medical Journal 85: 90­96. "Hyperglycaemic Nonketotic Coma as a Cause of Emergency Mosley, W. H., J. L. Bobadilla, and D. T. Jamison. 1993. "The Health Hyperglycaemic Admission in Baragwanath Hospital." South African Transition: Implications for Health Policy in Developing Countries." Medical Journal 85: 173­76. In Disease Control Priorities in Developing Countries, ed. D. T. Jamison, Rwiza, H. T., A. B. M. Swai, and D. G. McLarty. 1986. "Failure to Diagnose W. H. Mosley, A. R. Measham, and J. L. Bobadilla, 673­99. New York: Diabetic Ketoacidosis in Tanzania." Diabetic Medicine 3: 181­83. Oxford University Press. Setel, P., N. Unwin, K. Alberti, and Y. Hemed. 2000. "Cause-Specific Adult Motala, A. A., M. A. K. Omar, and F. J. Pirie. 2000. "Type 1 Diabetes Mortality: Evidence from Community-Based Surveillance, Selected Mellitus in Africa: Epidemiology and Pathogenesis." Diabetes Sites, Tanzania, 1992­1998." Morbidity and Mortality Weekly Report 49: International 10: 44­47. 416­19. Moukouri, D. T., C. McMoli, C. Nouedoui, and J. C. Mbanya. 1995. "Les Sobngwi, E., J. F. Gautier, and J. C. Mbanya. 2003. "Exercise and the aspects cliniques de la rétinopathie à Yaoundé." Médecine d'Afrique Prevention of Cardiovascular Events in Women." New England Journal Noire 42 (89): 423­28. of Medicine 348: 77­79. Murray, C. J. L., and A. D. Lopez. 1997. "Mortality by Cause for Eight Sobngwi, E., E. Mauvais-Jarvis, P. Vexiau, J. C. Mbanya, and J. F. Gautier. Regions of the World: Global Burden of Disease Study." Lancet 349: 2001. "Diabetes in Africans. Part 1: Epidemiology and Clinical 1269­76. Specificities." Diabetes and Metabolism 27: 628­34. Nambuya, A. P., M. A. Otim, H. Whitehead, D. Mulvaney, R. Kennedy, and Sobngwi, E., J-C. Mbanya, E. N. Moukouri, and K. B. Ngu. 1999. D. R. Hadden. 1996. "The Presentation of Newly-Diagnosed Diabetic "Microalbuminuria and Retinopathy in a Diabetic Population of Patients in Uganda." Quarterly Journal of Medicine 89: 705­11. Cameroon." Diabetes Research and Clinical Practice 44: 191­96. Niang, E. H., S. N. Diop, E. A. Sidibe, M. Badiane, J. R. Lamouche, and A. Sobngwi, E., J. C. Mbanya, N. C. Unwin, A. P. Kengne, L. Fezeu, E. M. M. Sow. 1994."Echographic and Velocimetric Aspects of Arteriopathies Minkoulou, T. J. Aspray, and K. G. Alberti. 2002. "Physical Activity and in the Diabetic." Dakar Medical 39: 37­42. Its Relationship with Obesity, Hypertension and Diabetes in Urban Nkegoum, A. V. 2002. "Coût direct et indirect du diabéte en l'absence de and Rural Cameroon." International Journal of Obesity and Related complications chroniques a Yaoundé, Cameroun." MD thesis, Metabolic Disorders 26: 1009­16. University of Yaoundé I, Cameroon. Swai, A. B. M., J. Lutale, and D. G. McLarty. 1990. "Diabetes in Tropical Ohwovoriole, A. E., J. A. Kuti, and S. I. Kabiawu. 1988. "Casual Blood Africa: A Prospective Study 1981­7: Characteristics of Newly Glucose Levels and Prevalence of Undiscovered Diabetes Mellitus in Presenting Patients in Dar es Salaam, Tanzania 1981­7." British Medical Lagos Metropolis Nigerians." Diabetes Research and Clinical Practice 4 Journal 300: 1103­7. (2): 153­58. ------. 1993. "Prospective Study of Incidence of Juvenile Diabetes Omar, M. A., M. A. Seedat, R. B. Dyer, A. A. Motala, L. T. Knight, and P. J. Mellitus over 10 Years in Dar es Salaam, Tanzania." British Medical Becker. 1994. "South African Indians Show a High Prevalence of Journal 306: 1570­72. NIDDM and Bimodality in Plasma Glucose Distribution Patterns." Swai, A. B. M., D. G. McLarty, F. Sherrif, L. M. Chuwa, E. Maro, Z. Diabetes Care 17: 70­74. Lukmanji, W. Kermali, W. MaKene, and K. G. Alberti. 1990. "Diabetes Omar, M. A. K., M. A. Seedat, A. A. Motala, R. B. Dyer, and P. Becker. 1993. and Impaired Glucose Tolerance in an Asian Community in Tanzania." "The Prevalence of Diabetes Mellitus and Impaired Glucose Tolerance Diabetes Research and Clinical Practice 8: 227­34. 286 | Jean-Claude Mbanya and Kaushik Ramiaya Teuscher, T., P. Baillod, J. B. Rosman, and A. Teuscher. 1987. "Absence Complications of Type II Diabetes Mellitus. Putting Evidence into of Diabetes in a Rural West African Population with a High Practice." Journal of General Internal Medicine 12: 567­80. Carbohydrate/Cassava Diet." Lancet (8536): 765­68. Walker, R. W., D. G. McLarty, H. M. Kitange, D. Whiting, G. Masuki, Tooke, J. E., S. Thomas, and G. C. Viberti. 2000. "Proteinuria in Diabetes." D. M. Mtasiwa, H. Machibya, N. Unwin, and K. G. Alberti. 2000. Journal of Royal College London 34: 336­39. "Stroke Mortality in Urban and Rural Tanzania." Lancet 355: 1684­87. Tuomilehto, J., J. Lindstrom, J. G. Eriksson, T. T. Valle, H. Hamalainen, Whiting, D. R., L. Hayes, and N. C. Unwin. 2003. "Challenges to Health P. Ilanne-Parikka, S. Keinanen-Kiukaanniemi, et al. 2001. "Prevention Care for Diabetes in Africa." Journal of Cardiovascular Risk 10: 103­10. of Type 2 Diabetes Mellitus by Changes in Lifestyle among Subjects WHO (World Health Organization). 1998. Population Ageing--A Public with Impaired Glucose Tolerance." New England Journal of Medicine Health Challenge. Geneva: WHO. 344: 1343­50. ------. 1999. "Diagnosis and Classification of Diabetes Mellitus." In UKPDS Group. 1998a. "Intensive Blood-Glucose Control with Definition, Diagnosis and Classification of Diabetes Mellitus and Its Sulphonylureas or Insulin Compared with Conventional Treatment Complications. Geneva: WHO. and Risk of Complications in Patients with Type 2 Diabetes: UKPDS ------. 2000. The World Health Report 2000--Health Systems: Improving 33." Lancet 352: 837­53. Performance. Geneva: WHO. ------. 1998b. "Tight Blood Pressure Control and Risk of Macrovascular ------. 2003. The World Health Report 2002--Reducing Risks, Promoting and Microvascular Complications in Type 2 Diabetes: UKPDS 38." Healthy Life. Geneva: WHO. British Medical Journal 317: 703­12. Wojtaszewski, J. F., B. F. Hansen, J. Gade, B. Kiens, J. F. Markuns, L. J. UNFPA (United Nations Population Fund). 2000. State of World Goodyear, and E. A. Richter. 2000. "Insulin Signaling and Insulin Population. New York: United Nations. Sensitivity after Exercise in Human Skeletal Muscle." Diabetes 49 (3): Unwin, N., F. Mugusi, T. Aspray, D. Whiting, R. Edwards, J. C. Mbanya, 325­31. et al. 1999. "Tackling the Emerging Pandemic of Non-Communicable Yudkin, J. S. 2000. "Insulin for the World's Poorest Countries." Lancet 355: Diseases in Sub-Saharan Africa: The Essential NCD Health 919­21. Intervention Project." Public Health 113: 141­46. Zouvanis, M., A. C. Pieterse, H. C. Seftel, and B. I. Joffe. 1997. "Clinical Vaughan, P., L. Gilson, and A. Mills. 1989. "Diabetes in Developing Characteristics and Outcome of Hyperglycaemic Emergencies in Countries: Its Importance for Public Health." Health Policy and Johannesburg Africans." Diabetic Medicine 14: 603­6. Planning 4: 97­109. Vijan, S., D. Stevens, W. Hermann, M. Funnell, and C. Standiford. 1997. "Screening, Prevention, Counseling, and Treatment for the Diabetes Mellitus | 287 Chapter 20 Cancers Freddy Sitas, Max Parkin, Zvavahera Chirenje, Lara Stein, Nokuzola Mqoqi, and Henry Wabinga Cancer has received low priority for health care services in age of 65 years is only about 20 percent lower than that of Sub-Saharan Africa. The reason is undoubtedly the over- her sisters in Western Europe (table 20.2). Yet the facilities whelming burden of communicable diseases, as illustrated for providing treatment for cancer cases in most of Africa by the proportions of deaths by major categories. Table 20.1 are minimal, as illustrated by the sparse distribution of shows estimates, for Africa and for the whole world, made radiation therapy services in Africa (figure 20.1) (Levin, by the World Health Organization (WHO) of the percent- El-Gueddari, and Meghzifene 1999). ages of deaths due to different causes in the year 2002 The noncommunicable diseases, such as cancers, are (WHO 2004). emerging health problems that need to be dealt with appro- From a purely objective point of view, therefore, con- priately to sustain public health advances that have already centration on health problems in Africa that have been been achieved. Increases in the prevalence of tobacco con- largely solved in the developed world (infant and child sumption and immunosuppression induced by the human mortality, maternal mortality, infectious diseases) appears immunodeficiency virus (HIV), coupled with such existing eminently reasonable. Unfortunately, these "old" diseases risk factors for cancer as alcohol; the high prevalence of coexist in Africa with the emergence of new ones, most cancer-associated infectious agents like human papillo- prominently the acquired immune deficiency syndrome maviruses (HPV), hepatitis B viruses (HBV), and human (AIDS), but also some of the noncommunicable diseases, herpesvirus-8 (HHV8); and environmental exposure to such as hypertension, diabetes, accidents and violence toxins, such as aflatoxins, will have an important impact on (Motala 2002; Reza, Mercy, and Krug 2001; Seedat 2000; future cancer patterns and incidence. Even despite declining Walker et al. 2000), and cancer. Cancer is not a rare disease overall life expectancy as a result of the HIV epidemic, in Africa. Even ignoring the huge load of AIDS-related Africans will continue to age, which will contribute to Kaposi's sarcoma, the probability that a woman living in cancer's becoming an increased burden on health services, present-day Kampala or Harare will develop a cancer by the both in relative and absolute terms. 289 Table 20.1 Estimated Percentages of Deaths, by Cause, 2002 Until quite recently, knowledge of cancer patterns was based primarily on clinical and pathological case series from Cause World Africa the 1950s and 1960s,which were the subject of several reviews No. of deaths (all causes) 57,029,000 10,664,000 that drew together information on the relative frequency of Infectious and parasitic diseases, 7.3 22.6 different types of cancer in different areas in order to piece excluding HIV/AIDS, respiratory infections together an overall picture (Clifford, Linsell, and Timms HIV/AIDS 4.9 19.6 1968; Cook and Burkitt 1971; Oettlé 1964). Statistics on dis- Respiratory infections 6.9 10.5 ease mortality are particularly sparse. Only about 0.25 per- Maternal and perinatal causes 5.2 7.4 cent of the population of Sub-Saharan Africa is covered by Cancer 12.5 3.8 accurate death registration systems. The countries that have Cardiovascular disease 29.3 9.7 reasonably accurate death registration include islands like Injuries, violence 9.1 7.0 Mauritius and the Seychelles, which are unlikely to be rep- Other causes 24.8 19.4 resentative of the region, and no country on the mainland of Sub-Saharan Africa has data of sufficient quality for the esti- Source: WHO 2004. mation of national mortality rates (Mathers et al. 2005). Table 20.2 Cumulative Incidence of Cancer in Women up to Hence, reliance has to be placed on indirect measures of 64 Years of Age, 1993­97 mortality and on the few cancer registries that do exist Country Cumulative incidence (%) across Africa, now covering roughly 8 percent (Parkin et al. England 15.2 2003) of this population. An exception to this has been Francea 14.0 South Africa, which until 1990 had almost complete death Sweden 14.4 notification for whites, mixed race "coloureds," and Asian Uganda (Kampala) 11.3 Indians, comprising about 20 percent of the population. Zimbabwe (Harare) 12.6 Population group identifiers on death notification forms were removed in 1991 but reintroduced in 1998. National Source: Parkin et al. 2003. Note: Kaposi's sarcoma and nonmelanoma skin cancer are excluded from table. coverage of deaths in South Africa across all populations has a. Based on data from nine cancer registries. now increased to over 90 percent (Dorrington et al. 2001). Figure 20.1 Distribution of Radiation Therapy Services Since the 1990s there has been a resurgence of interest in in Africa cancer incidence in Africa, and data from cancer registries from Sub-Saharan Africa have been published from West 30° 20° 10° 0° 10° 20° 30° 40° 50° IBRD 34531 MARCH 2006 Africa in The Gambia (Bah et al. 2001), Mali (Bayo et al. Mediterranean Sea 1990), Guinea (Koulibaly et al. 1997), and Côte d'Ivoire 30° 30° (Echimane et al. 2000). Data from East Africa are available Red from cancer registries in Kampala, Uganda (Wabinga et al. 20° Sea 20° 2000), and from southern Africa from the Zimbabwe Cancer Registry in Harare (Chokunonga et al. 2000), and the 10° 10° Malawi Cancer Registry in Blantyre (Banda et al. 2001). Cancer registration in economically underdeveloped populations, such as all the countries of Sub-Saharan Africa, 0° ATLANTIC 0° OCEAN INDIAN OCEAN is a difficult undertaking for a variety of reasons (Parkin et al. 10° 10° 2003). The major challenge is to ensure that all new cases of cancer are identified. Cases can be found only when they come into contact with health services: hospitals, health cen- 20° 20° less than 2 million/machine ters, clinics, and laboratories. When resources are restricted, 2 million ­ 10 million/machine more than 10 million/machine the proportion of the population with access to such insti- 30° no functional equipment 30° no data tutions may be limited, and the statistics generated will thus international boundaries not truly reflect the pattern of cancer. The ease with which This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. the cases can be identified also depends on the extent of 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° Source: Levin, El-Gueddari, and Meghzifene 1999. medical facilities available and the quality of statistical and 290 | Freddy Sitas, Max Parkin, Zvavahera Chirenje, Lara Stein, Nokuzola Mqoqi, and Henry Wabinga record systems already in place (for example, pathology areas reflects that in the country as a whole, given what is request forms, hospital discharge abstracts, treatment known of urban­rural differences in cancer patterns in records, and so forth). It is impossible to know, without an other areas of the world. extensive population survey, what proportion of those with The International Agency for Research on Cancer (IARC) cancer never come into contact with modern diagnostic or has published the available data on cancer incidence and treatment services, instead making use only of traditional other cancer data from a variety of sources (Parkin et al. healers or receiving no care at all. 2003). Such data have also been used to prepare a set of esti- In the past, studies have suggested that some sections of mates of incidence and mortality at the national level for the the population may have been underrepresented in hospital year 2002 (Ferlay et al. 2005). These sources are extensively statistics, particularly older women and young men, both of used in this chapter. We also draw upon the few available whom were more likely to return to their rural homes to series from which it is possible to make some inferences seek care (Flegg Mitchell 1966). However, currently, this about temporal trends in cancer incidence: the two cancer underrepresentation is probably rather rare in contempo- registries with data available in the 1960s--Kampala rary urban Africa. Most cancer patients will, eventually, seek (Uganda) and Ibadan (Nigeria)--and the mortality data medical assistance, although often at an advanced stage of sets from South Africa referred to earlier. disease. The situation in rural areas may be quite different, According to the 2002 estimates of cancer incidence for but almost all the present-day cancer registries are located in the Sub-Saharan Africa region, about half a million urban centers. From an epidemiological point of view, one (530,000) new cases of cancer occurred annually, 251,000 must guess at how well the cancer profile from the urban in males and 279,000 in females. Table 20.3 shows the Table 20.3 Estimated Number of New Cases and Age-Standardized (World) Incidence Rates for the Leading Cancers in Males and Females, 2002 (per 100,000 people) Eastern Middle Northern Southern Western Sub-Saharan Africa Africa Africa Africa Africa Africa Africa Sex/cancer Cases ASR Cases ASR Cases ASR Cases ASR Cases ASR Cases ASR Cases ASR Males Kaposi's sarcoma 23,094 23.0 10,097 30.0 194 0.3 3,065 13.2 3,740 4.6 39,995 16.4 40,190 12.3 Liver 14,012 21.1 7,744 27.8 2,351 4.2 1,072 7.0 10,637 15.3 33,465 18.6 35,812 14.8 Prostate 7,054 13.8 4,975 24.5 2,908 5.8 4,778 40.5 9,947 19.3 26,755 19.8 29,663 16.0 Esophagus 11,174 19.1 369 1.5 1,140 2.1 2,804 19.7 802 1.3 15,150 9.9 16,289 7.8 Non-Hodgkin's lymphoma 7,264 7.1 1,525 4.5 3,124 4.4 846 4.8 4,868 5.7 14,503 6.0 17,626 5.6 Stomach 4,687 7.4 3,283 13.4 2,550 4.4 1,183 8.2 2,131 3.4 11,287 6.9 13,836 6.2 Colon and rectum 4,019 6.1 627 2.3 3,150 5.1 1,553 11.3 3,430 5.1 9,630 5.6 12,778 5.4 All sites but skin 118,903 158.7 39,212 141.9 60,011 99.0 31,626 213.7 61,610 90.0 251,351 135.6 311,363 126.0 Females Cervix uteri 33,903 42.7 8,201 28 8,175 12.1 7,698 38.2 20,919 29.3 70,723 35.2 78,897 29.3 Breast 15,564 19.5 5,173 16.5 16,588 23.2 6,474 33.4 21,397 27.8 48,609 23.5 65,197 23.4 Kaposi's sarcoma 10,814 9.5 3,501 8.6 65 0.1 1,452 5.7 1,393 1.4 17,160 6.1 17,226 4.6 Liver 6,267 8.6 4,571 13.4 1,442 2.2 469 2.5 4,162 5.6 15,470 7.6 16,912 6.2 Stomach 3,883 5.5 3,780 12.6 1,671 2.5 686 3.7 2,331 3.6 10,680 5.7 12,350 4.9 Non-Hodgkin's lymphoma 4,741 4.4 2,249 6.9 1,830 2.4 614 3.0 2,995 3.5 10,598 4.4 12,428 3.9 Ovary, etc. 4,706 5.8 1,182 3.3 1,892 2.6 1,003 5.2 3,601 4.6 10,491 5.0 12,384 4.3 Colon and rectum 2,997 4.1 951 3.3 2,707 4.0 1,644 8.9 2,605 3.5 8,195 4.3 10,903 4.2 Esophagus 5,411 8.0 62 0.2 875 1.4 1,301 7.0 410 0.6 7,184 4.1 8,058 3.4 All sites but skin 129,029 156.7 38,857 121.5 59,603 85.2 32,170 163.2 78,740 104.0 278,797 133.4 338,397 121.0 Source: Ferlay et al. 2004. Note: ASR age-standardized rate. Cancers | 291 Figure 20.2 Major Cancer Types in Sub-Saharan Africa, Both CERVICAL CANCER Sexes, All Ages Cancer of the cervix is the leading cancer in women in Sub- liver 9.2% Kaposi`s sarcoma Saharan Africa with an estimated 70,700 new cases occurring 10.8% in 2002 (the total in the whole continent was 78,900 cases). breast 9.2% Estimated rates for eastern and southern Africa of 30 to 60 per 100,000 are higher than those found in the rest of Sub- cervix uteri Saharan Africa (20 to 35 per 100,000), but the reasons for this prostate 13.3% 5.0% difference are unclear. In many developed countries, such as the United Kingdom and Sweden, mortality from cancer of non-Hodgkin's lymphoma the cervix declined between the early 1900s and the 1960s 4.7% and then declined further as a result of the introduction of esophagus 4.2% national screening programs (Bergstrom, Sparen, and Adami 1999). However, in Bulawayo between 1963 and 1977 and in stomach 4.1% Kampala in the 1960s, 1970s, and 1990s, cancer of the cervix colon/rectum 3.4% has appeared to increase in incidence over time (Skinner et al. 1993; Wabinga et al. 2000). No increases over time were other 36.1% observed in Nigeria and South Africa (Parkin et al. 2003). Source: Adapted from Ferlay et al. 2004. It was noted early that cervical cancer has quite marked Note: Number of cases was 530,148. differences in incidence according to classical demographic variables (social class, marital status, ethnicity, religion). leading cancer types by region (including the northern Later, epidemiological studies (mainly case-control studies) Africa region) and by sex. Figure 20.2 shows the major can- showed a consistent association between risk and early age at cer types in Sub-Saharan Africa; overall, world-standardized initiation of sexual activity, increasing number of sexual cancer rates were estimated to be 133 per 100,000 females partners of females or of their sexual partners, and other and 136 per 100,000 males. indicators of sexual behavior. These findings were strongly suggestive of a causative role for a sexually transmitted agent. The top six cancers in males were the following: It is now recognized that certain sexually transmitted onco- · Kaposi's sarcoma (15.9 percent) genic human papillomaviruses constitute the necessary cause · liver (13.3 percent) of cervical cancer. However, additional independent risk · prostate (10.7 percent) factors include increasing number of pregnancies, exposure · esophagus (6.0 percent) to oral contraceptives, smoking, and specific dietary patterns. · non-Hodgkin's lymphoma (5.8 percent) At the onset of the AIDS epidemic, cancer of the cervix · stomach (4.5 percent). was classified as an AIDS-defining cancer by the U.S. Centers for Disease Control and Prevention (CDC 1993). In females, the following were the leading cancers: But it is far from clear that HIV infection really increases the · cervix (25.4 percent) risk of invasive cervical cancer. No change in cervical cancer · breast (17.4 percent) incidence has been demonstrated in some centers like · Kaposi's sarcoma (6.2 percent) Harare, where HIV/AIDS has been endemic for some time · liver (5.5 percent) (Chokunonga et al. 1999). In Kampala the increase in cervi- · stomach (3.8 percent) cal cancer incidence began before the advent of AIDS · non-Hodgkin's lymphoma (3.8 percent). (Wabinga et al. 2000). With respect to cervical intraepithe- lial neoplasia (CIN), most studies failed to adjust for the fact Each of these cancers is briefly discussed in this chapter. that, for obvious reasons, women infected by HIV were very In addition, tobacco-related cancers (especially lung cancer, often also infected by HPV (with a consequently high risk of which currently ranks seventh in males) and HIV-related CIN). Careful adjustment for such confounding suggests cancers (cancers aside from Kaposi's sarcoma) are discussed, that HIV has an independent effect on risk of CIN but that as these are likely to increase over time as both these it is small; there is an interaction between the effects of HIV epidemics mature. and HPV, as might be expected, if the role of HIV is indirect, 292 | Freddy Sitas, Max Parkin, Zvavahera Chirenje, Lara Stein, Nokuzola Mqoqi, and Henry Wabinga through creation of immune suppression and dysfunction (Chirenje et al. 2001), then this method may provide a use- (Mandelblatt et al. 1999). ful alternative to the conventional Pap test, not least in that Case-control and descriptive studies on cancer of the treatment is provided during the same visit as the screening cervix in Africa have shown associations of the disease test, thus dispensing with the requirement to recall women similar to those observed in Western countries with respect for diagnosis and therapy. to number of partners, level of education, high parity, and Vaccines against the leading HPV serotypes have now steroid contraceptives; however, genital hygiene, vaginal dis- been developed, and programs may be implemented for charge, alcohol, and male circumcision were also found in women before they become sexually active. However, it is certain studies to be important (Parkin et al. 2003). HIV was unclear how long the protection will last and whether the found to be associated with cervical cancer in case-control vaccine will also be effective in reducing the incidence of and cohort studies in South Africa and Uganda (Mbulaiteye cancer of the cervix among women who are infected. The et al., forthcoming; Newton et al. 2001; Sitas et al. 2000) with ongoing trials are expected to clarify such issues. As men are odds ratios between 1.6 and 2.4; however, such a weak asso- also carriers of HPV, future studies ought to measure any ciation could easily be due to confounding by sexual activity, added effectiveness of vaccination in this group. and other studies have shown no association (Newton et al. 1995; Sitas et al. 1997; ter Meulen et al. 1992). With regard to HPV, subtypes 16, 18, and 31 appear to be the leading ones, BREAST CANCER but other sexually transmitted infections causing chronic cervico-vaginal inflammation may increase the risk of cervi- Breast cancer is the second most common cancer among cal cancer. women in Sub-Saharan Africa, accounting for 16.8 percent Before the introduction of screening programs in the of all female cancers. Central, West, and East Africa appear 1960s and 1970s, the incidence in most of Europe, North to have lower incidence rates than southern Africa, the latter America, and Australia and New Zealand was much as we estimated at 33.4 per 100,000. An estimated total of 48,600 see it in Africa today: it was 38 per 100,000 in the Second cases occurred in Sub-Saharan Africa in 2002. National Cancer Survey of the United States, for example Worldwide, risk factors for female breast cancer include (Dorn and Cutler 1959). National screening programs have menstrual and reproductive factors, high body mass index been responsible for the further decline in the incidence of (BMI), family history of breast cancer, and certain genetic cancer of the cervix. Pap test screening, with coverage of mutations, including BRCA1/2. Other suggested risk factors over 80 percent of the female population over 35 years of age include, to a much lesser extent, high alcohol consumption, appears to be the most effective method in reducing the contraceptive use, and the use of certain postmenopausal incidence of cervical cancer. For example, if women were hormone replacement therapies. Reproductive and hor- offered screening three times in their lifetime (at about ages monal factors appear to be the most important, with risk 35, 45, and 55) the incidence of cancer of the cervix would being increased by early menarche, late menopause, late be halved (Miller 1992). age at first birth, and low parity (Henderson, Ross, and Given the complex organization of screening programs, Bernstein 1988). no organized national cervical cancer screening program Studies in Sub-Saharan Africa have also found reproduc- exists in Africa. Reasons for this include lack of good quality tive and hormonal factors to be important, reporting cytology services, difficulty of long-term follow-up in increased risk with advanced age at first pregnancy and many communities, lack of education, and lack of postal delivery, low parity, and late age at menarche (Adebamowo facilities and infrastructure. But many countries in Sub- and Adekunle 1999; Coogan et al. 1996; Shapiro et al. 2000; Saharan Africa do not have the ability to diagnose or treat Ssali, Gakwaya, and Katangole-Mbidde 1995). CIN. In other countries some attention has been given to the In Sub-Saharan Africa, higher incidence rates and relative value of screening by visual inspection after acetic acid frequencies of breast cancer have been reported in association impregnation of the cervix (University of Zimbabwe/ with urban than with rural residence (Oettlé and Higginson JHPIEGO Cervical Cancer Project 1999). The high negative 1966; Schonland and Bradshaw 1968), but data are sparse. predictive value of this approach suggests that few signifi- The incidence of breast cancer is much higher among white cant lesions will be missed. If appropriately and safely women in Africa than among black African women; for treated by effective, affordable methods like cryotherapy example, in Harare between 1993 and 1995, the incidence Cancers | 293 was 127.7 per 100,000 in whites and 20.4 in blacks About 1 percent of all breast cancer cases occur in men, (Chokunonga et al. 2000). These differences may be a reflec- with the male-to-female ratio being higher in black and tion of the distribution of lifestyle factors thought to be African populations than among white populations (Parkin important in the development of breast cancer, for example, et al. 2003; Sasco, Lowels, and Pasker de Jong 1993). low parity and high body mass. A review of the literature indicates a deficit of studies on Breast cancer risk has been associated with socioeconomic breast cancer risk in Sub-Saharan Africa, and further status, with women of higher social class (as measured by research could be beneficial. As certain groups become more education, income, housing, and so forth) having a higher Westernized and urbanized, with associated changes in diet, risk (Kogevinas et al. 1997). Once again, such differences are later childbirth, and reduced parity and periods of breast- most likely a reflection of different prevalences of risk factors feeding, breast cancer incidence may increase. Public health among social classes (for example, parity, age at menstrua- campaigns should encourage breastfeeding unless there are tion and menopause, height, weight, alcohol consumption). good reasons not to (for example, HIV-infected mothers The effect of oral contraceptive hormones on the risk of where milk powder and sterile water are freely available). breast cancer has been the subject of much research. There There is no organized mammography screening program in appears to be a small but detectable risk in women currently Sub-Saharan Africa. using oral contraceptives, but this diminishes when contra- ception ceases, and after 10 years, none of the excess risk remains (Reeves 1996). A case-control study in South Africa KAPOSI'S SARCOMA found that combined oral contraceptives may result in a small increase in risk, confined to women below the age of Prior to the HIV/AIDS era, Kaposi's sarcoma was a rare can- 25 years, but that injectable progesterone contraceptives did cer in Western countries, seen mainly among immigrants not increase risk (Shapiro et al. 2000). from the Mediterranean littoral and African regions and in Dietary fat appears to be correlated with the risk of breast immunosuppressed transplant recipients. Meanwhile, in cancer in interpopulation studies (Prentice and Sheppard Africa, the incidence of Kaposi's sarcoma varied 100-fold, 1990), but the association has been difficult to confirm in being most common in central and eastern Africa and rare studies of individuals (Hunter et al. 1996). However, obesity in northern and southern Africa (IARC 1996; Oettlé 1962); in postmenopausal women has been identified as a risk in certain parts of central and eastern Africa, Kaposi's factor in Europe (Bergstrom et al. 2001) as well as in Sub- sarcoma was as common as cancer of the colon was in the Saharan Africa (Adebamowo and Adekunle 1999; Walker West (Cook-Mozaffari et al. 1998). There appears to be et al. 1989). Although traditional diets in Africa are typically some geographical association with the prevalence of low in animal products, especially fat, and high in fiber human herpes virus-8, now regarded as a necessary cause (Labadarios et al. 1996; Manning et al. 1971), this pattern is for the development of Kaposi's sarcoma (Dukers and Rezza being modified by urbanization and Westernization of 2003). The incidence of Kaposi's sarcoma has increased over lifestyles, which may lead to an increase in breast cancer 1,000-fold in populations at high risk of HIV in some incidence in African populations. A case-control study in Western countries (Biggar et al. 1984; Rabkin, Biggar, and Cape Town did not find a protective effect of breastfeeding Horm 1991), but in the rest of the population the tumor still on breast cancer (Coogan et al. 1999). However, in a meta- remains relatively rare (Grulich, Beral, and Swerdlow 1992; analysis of 47 studies from 30 countries breastfeeding Rabkin, Biggar, and Horm 1991). In Africa, since the 1980s, appears to be protective; based on a reanalysis of about areas like Malawi, Swaziland, Uganda, and Zimbabwe, where 50,302 cases and 96,973 controls, two-thirds of the differ- Kaposi's sarcoma was relatively common before the era of ence in rates between developed and developing countries AIDS, the incidence of Kaposi's sarcoma has increased about were estimated to be attributed to breastfeeding (Inter- 20-fold, such that it is now the leading cancer in men and national Collaboration on HIV and Cancer 2002). the second leading cancer in women. In these cancer reg- At least part of the familial risk of breast cancer is medi- istries, overall age-standardized rates have increased by ated through the major susceptibility genes BRCA1 and about 15 percent, mainly as a result of HIV-associated BRCA2 (about 2 percent of breast cancer cases in Europe). Kaposi's sarcoma (for example, Bassett et al. 1995; Wabinga Very little is known of the prevalence of these mutations in et al. 1993; Wabinga et al. 2000). African populations, although family history of breast cancer According to the most recent estimates, 40,000 cases of is also a risk factor in this setting (Rosenberg et al. 2002). Kaposi's sarcoma in males and 17,200 cases in females were 294 | Freddy Sitas, Max Parkin, Zvavahera Chirenje, Lara Stein, Nokuzola Mqoqi, and Henry Wabinga estimated for 2002 for Sub-Saharan Africa; only 200 male Uganda. In Rwanda the age-standard incidence rate was and 65 female cases were estimated to occur in northern found to be 13 per 100,000 males and 15 per 100,000 Africa. The region most affected is central Africa (age- females (Newton et al. 1996). In western Uganda, stomach standardized rates in males of 30 per 100,000) followed by cancer was the second most common cancer, accounting for eastern, southern, and lastly western Africa, in line with the 12 percent of all male cancers and 6 percent of all female background prevalence of HIV in each of these regions. cancers (Wabinga et al. 2000). Bamako in Mali was another With regard to the effect of HIV infection, three case- area with a high incidence rate: 18.5 per 100,000 males and control studies from Africa showed increased risks of 30 to 15 per 100,000 females (Bayo et al. 1990). 50 in association with HIV, and these risks rise to 1,600 in There is evidence of a slight but not significant increase HIV-positive individuals with high HHV8 antibody titers in the incidence of stomach cancer over time in Kampala (Newton et al. 2002; Sitas et al. 1997; Sitas et al. 1999; Sitas (Wabinga et al. 2000). In Kivu Province of the Democratic et al. 2000). HHV8 in adults is associated with increasing Republic of Congo, the incidence rates of stomach cancer age, low educational standard, and increasing numbers of among males and females were 9 and 15 per 100,000, respec- sexual partners (Sitas et al. 1999). Antiretroviral therapy for tively, in 1956­60, but this dropped to 6 and 4.5 per 100,000 treating HIV in adults has caused a decline in the incidence in 1983­86 (Bourdeaux et al. 1988; Clemmensen, Maisin, of Kaposi's sarcoma in Western countries (International and Gigase 1962). However, in rural Kenya reported inci- Collaboration on HIV and Cancer 2000). HHV8 in children dence increased as a result of an endoscope acquired by the appears to be associated with infected mothers (Bourboulia main hospital there (McFarlane et al. 2001). Trend data from et al. 1998). In countries with a high prevalence of HIV, the rest of Africa are incomplete or inconsistent; however, Kaposi's sarcoma is now the leading cancer in children, in South Africa, between 1948 and 1964 no real change in causing almost a doubling in the childhood cancer inci- the relative frequency of stomach cancer was observed over dence (Chokunonga et al. 1999; Wabinga et al. 1993). time in one of the country's largest hospitals serving the pre- Antiretroviral drugs have now become more available in dominantly black population of Soweto, Johannesburg Botswana and recently in South Africa. If their use becomes (Robertson 1969), nor was a change observed in pathology- widespread, then a decline in the incidence of Kaposi's sar- based cancer national registrations between 1986 and 1995 coma would be expected; however, it is unclear whether (Sitas, Madhoo, and Wessie 1998). antiretrovirals (for example, zidovudine [AZT] or nevirap- Helicobacter pylori infection is now recognized as an ine) issued to mothers during delivery, which proved effec- important risk factor for cancer of the stomach (IARC tive in reducing mother-child transmission of HIV, would 1994); however, smoking and diets low in fruit and vegeta- cause a decline in Kaposi's sarcoma in children. bles and vitamin C, and high in salts appear to play an important role. Many studies have shown the prevalence of H. pylori in Africa to be about 80 percent and that infection STOMACH CANCER is acquired at a younger age than in Western countries (for example, Sathar et al. 1994). Chronic atrophic gastritis and A total of 13,800 cases of stomach cancer in males and intestinal metaplasia of the stomach are two key lesions 10,700 in females was estimated in Sub-Saharan Africa in in the natural history of stomach cancer. Very few studies in 2002. Age-standardized incidence rates in males varied, per Sub-Saharan Africa have measured the association between 100,000, from 3.4 in western Africa to 7.4 in eastern, 8.2 in gastric mucosal pathology and H. pylori. In summary, even southern, and 13.4 in central Africa. In western Africa, in a continent where the prevalence of H. pylori is high, where the incidence of stomach cancer is the lowest, the differences exist in the prevalence of H. pylori between those male-to-female ratio is 0.9 to 1; however, there is a male with a normal mucosa (0­33 percent) and those with gastri- predominance in all other areas (table 20.3). Despite the tis of any kind. Needless to say, the prevalence of gastritis generally low incidence rate in Africa, some populations (mild or moderate) is high, but the prevalence of severe or have a particularly high incidence rate. Clusters of high inci- chronic atrophic gastritis or intestinal metaplasia is low dence exist among the South African mixed race, or (Parkin et al. 2003). Two case-control studies from Africa coloured, population of 98 per 100,000. A high incidence show an association between H. pylori and stomach cancer, rate is also reported in the Great Lakes region that includes but the relative risks are low, probably because the mucosa Burundi, Kivu Province of the Democratic Republic of of patients with gastric cancer is unfavorable to the sur- Congo, Rwanda, northwestern Tanzania, and southwestern vival of H. pylori (Jaskiewicz et al. 1989; Louw et al. 2001). Cancers | 295 CagA positive strains, usually associated with more severe Table 20.4 Prevalence of Hepatitis C Virus IgG Antibodies gastric pathology and outcomes, are the predominant in Sub-Saharan Africa, 2000 strains in Africa (Ally et al. 1998), but their role in gastric Region Prevalence of HCV (%) carcinogenesis is unclear. Certain vacA genotypes appear to Eastern Africa 2.7 be more common in patients with gastric cancer (Kidd et al. Middle Africa 6.9 1999) and seem to be independent risk factors for the dis- Southern Africa 0.1 ease; however, no studies have been done in Sub-Saharan Western Africa 2.4 Africa on the relation between stomach cancer, host suscep- All Africa 3.0 tibility (in relation to inflammatory cytokines), and the other risk factors known to be associated with stomach can- Source: Madhava, Burgess, and Drucker 2002. cer (for example, diet, salt, smoking, and pickled foods) (see, for example, Coggon et al. 1989). There are many places in Chronic carriage of HBV or hepatitis C (HCV), causing Africa where food is salted or pickled to aid preservation, cirrhosis, or chronic hepatitis is the leading risk factor for but the relative importance of these risk factors in local liver cancer. The prevalence of HCV in Sub-Saharan Africa settings is unknown. varies between 6.9 percent in central Africa to 0.1 percent in southern Africa (table 20.4). HCV transmission is probably via blood transfusion, unsterile medical and dental proce- LIVER CANCER dures, and traditional practices, such as scarification; sexual transmission is thought to be rare (Madhava, Burgess, and Early observations in Africa have always noted the high Drucker 2002). occurrence of liver cancer (for example, Oettlé 1964), and it Persistence of the HBV surface antigen (HbsAg) in blood is still one of the leading cancer types in men and women, is an indicator of chronic carriage of HBV infection. The although the relative frequency has been reduced in conse- risk of liver cancer in persons with chronic HBV infection, quence of the large increase in the number of cases of as indicated by the detection of HbsAg in serum, ranges Kaposi's sarcoma resulting from the epidemic of HIV/AIDS. from 6- to 20-fold in different studies, and it is estimated Liver cancer is now the second leading cancer in men in Sub- that about two-thirds of liver cancer in Africa is attributed Saharan Africa and the fourth leading cancer in women to HBV (Pisani et al. 1997). Prevalence rates in Africa are (table 20.3). There were an estimated total of 33,500 cases in over 10 percent in central, western, and eastern Africa and males and 15,500 cases in females in 2002. Areas of high liver between 5 and 10 percent in southern Africa (Parkin et al. cancer incidence (mainly hepatocellular cancers) include 2003). countries like The Gambia, Guinea, and Senegal in West There are relatively few African studies on the risk of Africa, where liver cancers comprise a quarter or more of all HCV infection on the development of liver cancer. Those cancer cases, with incidence rates ranging from 30 to 50 per that have been conducted give relative risks ranging from 1.1 100,000 in men and 12 to 20 per 100,000 in women. Similarly, to 62 (Parkin et al. 2003). One study (Kirk et al. 2004) in central Africa, liver cancer is the leading cancer in Rwanda observed that, as has been found elsewhere, the risk of and in the Republic of Congo (Brazzaville); the estimated chronic infection by HCV and HBV is additive, suggesting rate is 15.4 per 100,000 for men and 8.9 per 100,000 for common mechanisms of carcinogenesis. women. Mozambique is reported to have high incidence Aflatoxin B1 (AFB1) is produced by molds of Aspergillus rates, although the only data are old (Prates and Torres 1965). sp. that are common contaminants of poorly stored grains. Few places in Sub-Saharan Africa have information on AFB1 is a known liver carcinogen of animals and humans cancer trends over time. In Ibadan, Nigeria, between (IARC 1993, 2002). In Sub-Saharan Africa, high levels of 1960­69 and 1998­99, there appears to be no change in inci- AFB1 contamination are found in groundnuts and, to a dence, whereas in Kampala, Uganda, between the 1960s and lesser extent, corn. Contamination of groundnuts by AFB1 the 1990s there appears to be a decline of liver cancer in men is quite widespread and frequently exceeds thresholds but not in women. However, a decline was noted in liver permitted in exports to most developed countries. Several cancer incidence between the 1970s and the 1980s among geographical studies have demonstrated correlations Mozambican miners working in South Africa (Harington, between AFB1 levels and the incidence of hepatocellular Bradshaw, and McGlashan 1983). cancer (see Parkin et al. 2003). 296 | Freddy Sitas, Max Parkin, Zvavahera Chirenje, Lara Stein, Nokuzola Mqoqi, and Henry Wabinga Iron overload, derived from food and drink preparation the highest rates (40.5 per 100,000). Rates of histologically in iron vessels, is a common condition in rural Africa, and diagnosed prostate cancer in South Africa are 40.1 per there have been several observations that elevated serum 100,000 in whites versus 14 per 100,000 in blacks, although ferritin levels are associated with liver cancer. In one small for blacks, access to diagnostic facilities has been limited case-control study in South Africa (Mandishona et al. 1998), (Parkin et al. 2003). In Zimbabwe (defined as being part of liver cancer cases had higher iron overload levels than eastern Africa), rates for whites and blacks were 70 versus 25 controls, corresponding to an odds ratio of 10.6 to 4.1 per 100,000 (Parkin et al. 2003). Central Africa follows with (depending on the control group used). rates of 24.5 per 100,000. Surprisingly, in West Africa, where Smoking, oral contraception, and alcohol consumption the majority of African-American men originated, the inci- (IARC 2004, 1999, and 1988, respectively) were also found dence rate of prostate cancer was estimated as 19.3 per to be important risk factors for liver cancer. This associa- 100,000 in 2002, compared with about 125 per 100,000 in tion, however, has not been extensively examined in Africa. the United States (Ferlay et al. 2005). High rates are observed Early vaccine trials against HBV suggest that 70 to 75 per- in other places with populations that are descended cent of chronic infections could be prevented. A random- from West Africa (for example, the Bahamas, Barbados, ized trial to measure the effectiveness of HBV vaccination in Trinidad). the prevention of liver cancer is under way in The Gambia, Histology of the prostate in elderly men often reveals but it will take many years before results are available. In latent malignant cells, and clearly, advances in diagnostic Taiwan, however, children born after the introduction of and screening methods can cause artificial increases in mass vaccination had a fourfold lower incidence than those reporting. This is illustrated by a fourfold increase in the born before its introduction (Chang et al. 1997). According incidence of histologically verified prostate cancer among to the WHO Web site, by 2002, about a dozen countries whites in South Africa (most whites were covered by private in Sub-Saharan Africa had introduced hepatitis B vaccine health insurance) compared with no change in incidence in into their infant immunization system (http://www.who. blacks between 1986 and 1995 (Sitas, Madhoo, and Wessie int/vaccines-surveillance/graphics/htmls/HepBvaccine 1998). Notably, in Cape Town in the 1950s prostate cancer UseMar02.htm). appeared to be more common in blacks than in whites Aflatoxin consumption could be reduced by improved (Muir-Grieve 1960). Increases over time have also been education of individuals and farmers by, for example, agri- noted in Kampala and in Ibadan, but it is unclear how much cultural extension officers. A trial in western Africa has of these increases represents a greater risk and how much shown that improved post-harvest storage of groundnuts can be attributed to increased awareness or a greater readi- can significantly reduce aflatoxin exposure in rural popula- ness to perform prostatectomy for urinary symptoms in tions (Turner et al. 2005). The public could be educated to elderly men (Parkin et al. 2003). avoid contaminated peanuts sold by vendors (Wild and Hall The consumption of fat and red meat has been implicat- 2000). Companies manufacturing peanut butter could be ed as a risk factor for prostate cancer in studies in developed better controlled by accepting peanuts only from certified countries, even though adjustment for total caloric intake farmers and by the testing of their products by independent was not always done. Associations with vegetable consump- regulatory authorities. tion have been inconclusive. Associations with anthropo- metric measures or a link with obesity have been inconclu- sive, and so have associations with numbers of sexual PROSTATE CANCER partners and history of sexually transmitted diseases, or STDs (Hayes et al. 2000; Key 1995; Kolonel 1996). In one For the year 2002, a total of 26,800 cases of prostate cancer case-control study from South Africa, prostate cancer was were estimated, comprising 10.6 percent of cancers of men associated with high intake of fat, meat, and eggs; eating out in Sub-Saharan Africa (Ferlay et al. 2005). The relatively of the house; and a low consumption of vegetables (Walker high incidence (and mortality) recorded in African popula- et al. 1992). tions is reflected in populations of African descent else- Sex hormones, modulated by polymorphisms on the where. Thus, within the United States, the black population long arm of chromosome X, play an important role in the has the highest incidence (and mortality) rates, some 72 per- development of prostate cancer (for example, Ross et al. cent higher than whites. Southern Africa appears to have 1998; Shibata and Whittemore 1997). Polymorphisms on Cancers | 297 the androgen receptor gene may vary by ethnic group and Africa, however, the association between HIV and non- may provide some explanation for the geographic variation Hodgkin's lymphoma has been in the region of 2.3 to 12.3 observed. However, no studies have been done on intereth- (Mbulaiteye et al., forthcoming; Newton et al. 2001; Parkin nic variations in androgen receptor polymorphisms in et al. 2003; Sitas et al. 1997; Sitas et al. 2000). The reason for Africa. the discrepancy in the association between HIV and non- Hodgkin's lymphoma between developed countries and Africa is unclear. Non-Hodgkin's lymphomas were increas- NON-HODGKIN'S LYMPHOMA ing in incidence in Western populations before the advent of HIV but have increased dramatically in high-risk groups The non-Hodgkin's lymphomas are composed of an affected by HIV (see, for example, Schultz, Boshoff, and extremely heterogeneous group of lymphoproliferative Weiss 1996). In Harare, Zimbabwe (Chokunonga et al. malignancies displaying distinct behavioral, prognostic, and 1999), and in Kampala, Uganda, there is now evidence of an epidemiological characteristics. Advances in molecular biol- increase in incidence between earlier cancer registration ogy, genetics, and immunology have resulted in extensive periods and periods in the 1990s (Parkin et al. 1999; Parkin changes in the classification of lymphoid tumors in the last et al. 2003). few decades. The WHO classifies tumors according to cell Burkitt's lymphoma affects mainly children between the lineage defined by immunophenotype (Jaffe et al. 2001). ages of five and nine. The jaw is affected 50 to 60 percent of Three broad categories are now recognized: B-cell neo- the time. Burkitt's lymphoma shows a peculiar geographic plasms, T/NK-cell neoplasms, and Hodgkin's lymphoma. distribution and has been reviewed by others (for example, Lymphocytic leukemias fall within the B-cell neoplasm Burkitt 1969; Williams et al. 1978; Wright 1973). It accounts group. for about a quarter to a half of childhood cancers in the east- A total of 14,500 cases in males (5.8 percent of all can- ern and central parts of Africa and in tropical West Africa, cers) and 10,600 cases in females (3.8 percent of all female and less frequently in other places. Burkitt identified a strik- cancers) were estimated for 2002 in Sub-Saharan Africa. In ing distribution 15 degrees north and south of the equator, most African populations non-Hodgkin's lymphoma is rel- with a southern tail into Mozambique. But even within this atively rare, but the relative frequency is above the world area Burkitt's lymphoma was rarer in higher altitudes. The average in North and Sub-Saharan Africa because of the areas where it was most common were typified by rainfalls high incidence of Burkitt's lymphoma in children in the over 50 centimeters per year and an average of the coolest tropical zone of Africa. As in Western countries, most non- month of greater than 15.6°C, which seem associated with Hodgkin's lymphomas in Africa are of B-cell type. In adults, the distribution of malaria endemicity (Burkitt 1969; clinical series show an excess of high-grade lymphomas and O'Conor 1970). Low socioeconomic status, family cluster- a deficit of nodular lymphomas. ing, and proximity to the plant species Euphorbia tirucalli Human T-cell lymphotrophic viruses (for example, have been suggested as important factors in the etiology of HTLV-I) are common in tropical Africa (IARC 1996) and Burkitt's lymphoma; however, the leading agent has been are a cause of T-cell lymphomas; however, the incidence of infection with Epstein-Barr virus (IARC 1997). these in Africa is low. Although Epstein-Barr virus DNA In a follow-up study of 42,000 children, those who devel- may be found in a small proportion of lymphomas, its role oped Burkitt's lymphoma had higher titers of antiviral in causing non-Hodgkin's lymphomas is unclear (IARC capsid antigen than in matched controls (de Thé et al. 1978; 1997). HCV infection has been implicated in B-cell non- Geser et al. 1982). The link with malaria appears to be a Hodgkin's lymphomas in some studies; the postulated result of the loss of cytotoxic T-cell control due to dysfunc- mechanism being through the stimulation of polyclonal tion of a subset of CD4 cells responsible for the induction of proliferation of B cells (reviewed by Parkin et al. 2003). HIV suppressor-cytotoxic CD8 cells. This may result in uncon- infection has been associated with 60-fold increased risks of trolled proliferation of B cells containing the Epstein-Barr developing non-Hodgkin's lymphomas in Western countries virus and resultant malignant transformation (for example, (for example, Beral et al. 1991); approximately 5 to 10 per- Pagano et al. 1992; Whittle et al. 1990). In a five-year period cent of HIV-infected persons will develop a lymphoma, and of malaria suppression (when chloroquine was issued to non-Hodgkin's lymphoma is the AIDS-defining illness in children under 10), Burkitt's lymphoma appeared to decline about 3 percent of HIV-infected patients (Remick 1995). In in incidence. Incidence returned to the original level after 298 | Freddy Sitas, Max Parkin, Zvavahera Chirenje, Lara Stein, Nokuzola Mqoqi, and Henry Wabinga the five-year program was completed (Geser, Brubaker, and Eastern Cape Province has been associated with the monot- Draper 1989). Burkitt's lymphoma is much rarer in adults, onous consumption of corn, which contains low levels of although Burkitt-like (or high-grade Burkitt-like) lym- niacin, riboflavin, vitamin C, zinc, calcium, and magnesium phomas appear to be occurring with increased frequency as (Van Rensburg 1981) and is sometimes contaminated with a result of HIV (Sitas et al. 2000). fungal toxins produced by Fusarium spp. Certain studies in A prevention program for non-Hodgkin's lymphomas this region have shown a geographical association with the can be carried out only after the taxonomy and causes are presence of Fusarium moliniforme, a common fungal con- further elucidated. It appears that antimalarial programs taminant of poorly stored corn. Other risk factors reported may have a significant impact on Burkitt's lymphoma in in the Transkei include infections with Candida albicans and children, and as in Western countries, widespread antiretro- the consumption of a green, leafy plant weed, Solanum viral therapy of HIV-positive individuals would cause a nigrum (Sammon 1992). decline in the incidence of non-Hodgkin's lymphoma. OTHER HIV-ASSOCIATED CANCERS CANCER OF THE ESOPHAGUS Aside from Kaposi's sarcomas and non-Hodgkin's lym- In 2002 a total of 15,150 cases of cancer of the esophagus phomas, other cancers that appear to be associated with HIV were estimated to occur in males in Sub-Saharan Africa and immune suppression are cancer of the conjunctiva and pos- 7,200 cases in females. Cancer of the esophagus shows a sibly cancers of the cervix, vulva or vagina, anus, and liver remarkable geographic distribution, being one of the lead- (IARC 1996). However, except for conjunctival cancers, the ing cancers in southern and East Africa (average incidence data, at least from Sub-Saharan Africa, are not yet conclusive in males about 19 per 100,000) but rare in West Africa (1 to (IARC 1996; Parkin et al. 2003). Conjunctival cancers are 2 cases per 100,000). Certain areas of high risk have been increasing in incidence in Malawi (Banda et al. 2001);Uganda reported from Kenya and the former Transkei homeland in (Newton et al. 2001; Parkin et al. 1999), and Zimbabwe the Eastern Cape Province of South Africa, where incidence (Chokunonga et al. 2000); these countries have some of the rates as high as 76.6 per 100,000 in males and 36.5 per most prolonged and highest levels of HIV prevalence in 100,000 in females were reported between 1991 and 1995 Africa, and it is anticipated that these cancers will increase in (Somdyala et al. 2003). Several studies between the 1950s time in other places in Africa that are affected by HIV. and the 1990s in South Africa, Uganda, and Zimbabwe have demonstrated that cancer of the esophagus has increased in incidence. But the latest available data from cancer registries TOBACCO-RELATED CANCERS in these countries show a declining trend in esophageal cancer incidence, particularly in males after 1990 (Parkin Tobacco smoking is by far the most important cause of lung et al. 2003; Somdyala et al. 2003). cancer. The evidence has been reviewed many times (IARC Tobacco and alcohol consumption, known risk factors 1986, 2004). In 1985 it was estimated that about 76 percent of for the development of esophageal cancer in many coun- all lung cancer worldwide (84 percent of cases in men and tries, have also been documented as important in Africa in 46 percent in women) could be attributed to tobacco smok- studies conducted from the 1980s onward; earlier studies ing (Parkin et al. 1994). However, in Africa, because smok- found no such association, probably because of the low ing is a relatively recent habit in most areas, the proportion alcohol concentration of noncommercial drinks. The net of tobacco-attributed lung cancers is low. effect of increasing commercial alcohol consumption, Only where the smoking habit has been established in a combined with increases in some places of tobacco con- significant percentage of the population for a prolonged sumption, on esophageal cancer trends is to date unclear. period of time is the proportion of tobacco-attributable There is no consistent evidence of an effect of homemade cancers also significant--85 percent of cases in males in brews and esophageal cancer risk in Africa. certain southern African populations and 68 percent in Esophageal cancer also appears to occur in areas of northern Africa, for example (Parkin and Sasco 1993). extreme poverty and poor nutritional status. The high inci- Because of the lower incidence of lung cancer in Africa dence of esophageal cancer in the Transkei region of the (and the low prevalence of tobacco consumption in most Cancers | 299 Figure 20.3 Number of Cigarettes Consumed per Adult cancer if they smoked 15 cigarettes or more per day was 13, but in Soweto the relative risk for smoking the same number 30° 20° 10° 0° 10° 20° 30° 40° 50° IBRD 34532 MARCH 2006 of cigarettes was 20.7. The latter relative risk is comparable Mediterranean TUNISIA 1,548 Sea MOROCCO 811 to that observed in some developed countries. 30° Canary Is. 30° (Sp) ALGERIA ARAB In a study from South Africa, which uses the death regis- 1,021 LIBYA REP. OF FORMER 1,378 EGYPT SPANISH 1,275 SAHARA Red tration system to ask the next of kin about the smoking 20° MAURITANIA CAPE Sea 20° status of the deceased, 61 percent of male and 48 percent of 317 VERDE MALI 223 NIGER ERITREA SENEGAL 803 CHAD SUDAN female deaths due to lung cancer were found to be attrib- THE GAMBIA 160 73 171 BURKINA DJIBOUTI 221 FASO GUINEA GUINEA-BISSAU 10° NIGERIA 10° 90 uted to smoking (compared with 80 to 90 percent in men 189 ETHIOPIA SIERRA LEONE CÔTE BENIN 457 TOGO CENTRAL D'IVOIREGHANA LIBERIA 580 AFRICAN REPUBLIC 96 165 329 491 SOMALIA CAMEROON 653 and 30 to 70 percent in women in Western countries (Parkin EQUATORIAL GUINEA UGANDA 180 SÃO TOMÉ AND PRÍNCIPE KENYA and Sasco 1993). It was estimated that in all about 22,000 0° GABON CONGO 432 0° 519 476 DEMOCRATIC REPUBLIC RWANDA 135 INDIAN OF CONGO BURUNDI 86 adult deaths (8 percent of total deaths compared with 15 Cabinda 138 OCEAN (ANGOLA) TANZANIA 177 ATLANTIC SEYCHELLES 10° percent of deaths in Western countries; Peto et al. 2004) 10° OCEAN COMOROS ANGOLA MALAWI Mayotte 447 were attributed to tobacco (Sitas et al. 2004). Surprisingly, ZAMBIA 123 (Fr) 385 MOZAMBIQUE more deaths from chronic obstructive pulmonary disease ZIMBABWE MADAGASCAR 432 315 20° 399 20° NAMIBIA BOTSWANA MAURITIUS and tuberculosis would be expected from tobacco than 1,554 SWAZILAND 165 number of cigarettes consumed deaths from lung cancer. The reason for the lower propor- per capita (adults) SOUTH LESOTHO 30° 30° AFRICA international boundaries 1,322 tions of lung cancers attributed to smoking is that some of such cancers can be attributed to occupational exposures, This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. 30° 20° 10° 0° 10° 20° 30° 40° 50° 60° environmental tobacco smoke, air pollution, and radon gas Source: http://apps.nccd.cdc.gov/nations/African_region.asp. exposure (Parkin and Sasco 1993). Other tobacco-attributed cancers studied in Africa include those of the bladder, places in Africa) there is a widespread misconception that cervix, larynx, and esophagus and oral cancers (summarized the hazards of tobacco are only relevant in developed coun- in chapters in Parkin et al. 2003). tries. However, it appears that tobacco consumption, partic- ularly of manufactured cigarettes, is increasing in Africa. Figure 20.3 shows the distribution of per capita consump- CONCLUSION tion of cigarettes in Africa in countries where data exist. It is notable that aside from southern and northern Africa, con- Over the past century, until about the 1980s (prior to the sumption is low. Typical per capita consumption in the advent of HIV/AIDS), the average age of most populations in United States, for example, is 2,255 cigarettes, and in China, Sub-Saharan Africa has increased because of improvements 1,791, per year. In a WHO survey it was found that between in the rates of both infant and adult mortality (Timaeus two decades, 1970­72 and 1990­92, 15 countries in Africa 1999). Since cancer risk is strongly related to age, the aging increased their consumption of cigarettes, 6 decreased, and population has experienced an increase in the numbers of 5 remained unchanged (WHO 1997). Data were unavailable cancers and in crude incidence. Cancer has therefore been for the rest of Africa. Adult smoking rates also vary signifi- an emerging public health problem. The HIV epidemic has cantly; prevalence in Africa among men varies from 10 to arguably caused the biggest change in cancer patterns, with 50 percent, and among women from 1 to 10 percent (WHO Kaposi's sarcoma now being the leading cancer type in men 1997). An exception may be the mixed-race population of and the third most common cancer in women. But also, cer- South Africa, where there has been a high prevalence tain cancer types, such as cancer of the lung, breast, prostate, (currently 40 to 50 percent) of smoking among women, and esophagus, have increased significantly as a result of and, indeed, lung cancer rates (and rates for other tobacco- changing lifestyles and changes in exposures to common associated cancers such as oral and esophageal cancers) are carcinogens. higher in southern Africa than the rest of Africa. Although the relative importance of many important Evidence of the recent effect of the tobacco epidemic in carcinogens has been described for many cancers in most Africa comes from the Northern Province (mainly rural) of Western countries, little is known about the distribution of South Africa (Mzileni et al. 1999) and Soweto (urban South these and the relative importance of the major causes of Africa) (Pacella-Norman et al. 2002). In the Northern cancer in Africa. Even in places with existing cancer reg- Province, the relative risk that males would develop lung istries, or well-resourced countries like South Africa, very 300 | Freddy Sitas, Max Parkin, Zvavahera Chirenje, Lara Stein, Nokuzola Mqoqi, and Henry Wabinga few cancers or common carcinogenic exposures are being Bah, E., D. M. Parkin, A. J. Hall, A. D. Jack, and H. Whittle. 2001. "Cancer in The Gambia: 1988­1997." British Journal of Cancer 84: researched in a systematic fashion, and there is therefore 1707­24. wide uncertainty about their relative importance and their Banda, L. T., D. M. Parkin, C. P. Dzamalala, and N. G. Liomba. 2001. evolution over time. The relative importance of cancers and "Cancer Incidence in Blantyre, Malawi 1994­1998." Tropical Medicine related exposures to them needs to be carefully assessed in and International Health 6 (4): 296­304. order to formulate appropriate health promotion strategies. Bassett, M. T., E. Chokunonga, B. Mauchaza, L. Levy, J. Ferlay, and D. M. Parkin. 1995. "Cancer in the African Population of Harare, Zimbabwe Given the tremendous variation in the genetics, lifestyle in 1990­92." International Journal of Cancer 63: 29­36. characteristics, and cancer patterns throughout Africa, it Bayo, S., D. M. Parkin, A. K. Koumare, A. N. Diallo, T. Ba, S. Soumare, and may be misleading to extrapolate cancer patterns from one S. Sangare. 1990. "Cancer in Mali, 1987­1988." International Journal of Cancer 45 (4): 679­84. area to the next, so better data from population-based can- Beral, V., T. Peterman, R. Berkelman, and E. S. Jaffe. 1991. "AIDS- cer registries and from mortality statistics are needed to Associated Non-Hodgkin Lymphoma." Lancet 337: 805­9. provide data of local relevance. Bergstrom, A., P. Pisani, V. Tenet, A. Wolk, and H. O. Adami. 2001. There have, however, been some positive developments. "Overweight as an Avoidable Cause of Cancer in Europe." International Journal of Cancer 91 (3): 421­30. Compared with 1978­82, when no data from population- Bergstrom, R., P. Sparen, and H. O. Adami. 1999. "Trends in Cancer of the based cancer registries in Sub-Saharan Africa existed (Muir Cervix Uteri in Sweden Following Cytological Screening." British et al. 1987), the information derived from cancer registries Journal of Cancer 81 (1): 159­66. in Sub-Saharan Africa now covers 8 percent of the popula- Biggar, R. J., J. Horm, J. F. Fraumeni, M. H. Greene, and J. J. Goedert. 1984. "Incidence of Kaposi Sarcoma and Mycosis Fungoides in the United tion. Despite this great achievement most of these registries States Including Puerto Rico, 1973-1981." Journal of the National are staffed by a part-time director and one or two clerks, Cancer Institute 73: 89­94. who do not know whether sufficient support will be forth- Bourboulia, D., D. Whitby, C. Boshoff, R. Newton, V. Beral, H. Carrara, coming from their Departments of Health or other poten- A. Lane, and F. Sitas. 1998. "Serologic Evidence for Mother to Child Transmission of Kaposi Sarcoma Associated Herpes Virus Infection." tial stakeholders. Given the short time that these registries Journal of the American Medical Association 280 (1): 31­32. have existed in Africa the impact of these on cancer inci- Bourdeaux, L., F. Renard, P. L. Gigase, Mukolo-Ndjolo, P. Maldague, and dence is still difficult to quantify. Yet long-term surveillance A. De Muynck. 1988. L'incidence des cancers à l'hôpital de Katana, Kivu, Est Zaire, de 1983 à 1986. Ann. Soc. Belg. Med Trop 68: 141­56. is necessary to quantify the impact of the epidemics of Burkitt, D. P. 1969. "Etiology of Burkitt's Lymphoma--An Alternative tobacco and AIDS and to evaluate the efficacy of cancer Hypothesis to a Vectored Virus." Journal of the National Cancer Institute control measures. 42: 19­28. Still, despite the dramatic reduction in life expectancy in Centers for Disease Control and Prevention (CDC). 1993. "Revised Classification System for HIV Infection and Expanded Surveillance many populations in Sub-Saharan Africa, age-standardized Case Definition for AIDS among Adolescents and Adults." Journal of cancer incidence in these registries has remained the same, the American Medical Association 269: 729­30. and in places where HIV prevalence is high, the overall inci- Chang, M. H., C. J. Chen, M. S. Lai, H. M. Hsu, T. C. Wu, M. S. Kong, dence of cancer seems to have increased by up to 15 percent. D. C. Liang, W. Y. Shau, and D. S. Chen. 1997. "Universal Hepatitis B Vaccination in Taiwan and the Incidence of Hepatocellular Carcinoma This is counterintuitive to the common belief that a reduc- in Children." Taiwan Childhood Hepatoma Study Group. New England tion in life expectancy due to HIV would cause a decline in Journal of Medicine 336: 1855­59. chronic disease. Chirenje, M., V. Rusakaniko, M. Akino, and Z. Mlingo. 2001. "A Randomised Clinical Trial of Loop Electrosurgical Excision Procedure (LEEP) versus Cryotherapy in the Treatment of Cervical Intraepithelial Neoplasia." Journal of Obstetrics and Gynecology 21 (6): 617­21. ACKNOWLEDGMENTS Chokunonga, E., L. Levy, M. Bassett, M. Z. Borok, B. G. Mauchazo, M. Z. Chirenje, and D. M. Parkin. 1999. "AIDS and Cancer in Africa. The The authors are grateful to Ruth Lawrence for preparing the Evolving Epidemic in Zimbabwe." AIDS 13: 2583­88. Chokunonga, E., L. M. Levy, M. T. Bassett, B. G. Mauchaza, D. B. Thomas, manuscript. and D. M. Parkin. 2000."Cancer Incidence in the African Population of Harare, Zimbabwe: Second Results from the Cancer Registry 1993­1995." International Journal of Cancer 85 (1): 54­59. REFERENCES Clemmensen, J., J. Maisin, and P. Gigase. 1962. "Preliminary Report on Cancer in Kivu and Rwanda-Urundi." University of Louvain, Institut Adebamowo, C. A., and O. O. Adekunle. 1999. "Case-Controlled Study of de Cancer, Louvain. the Epidemiological Risk Factors for Breast Cancer in Nigeria." British Clifford, P., C. A. Linsell, and G. L. Timms, eds. 1968. Cancer in Africa. Journal of Surgery 86 (5): 665­68. Nairobi: East African Publishing House. Ally, R., M. Hale, C. Hadjinicolaou, H. E. M. Sonnendecker, K. D. Bardhan, Coggon, D., D. J. Barker, R. B. Cole, and M. Nelson. 1989."Stomach Cancer et al. 1998. "Helicobacter Pylori in Soweto South Africa. CagA Status and Food Storage." Journal of the National Cancer Institute 81 (15): and Histopathology in Children." Abstract. Gastroenterology 114: A54. 1178­82. Cancers | 301 Coogan, P. F., R. W. Clapp, P. A. Newcomb, R. Mittendorf, G. Bogdan, J. A. ------. 1993. Some Naturally Occurring Substances: Food Items and Baron, and M. P. Longnecker. 1996. "Variation in Female Breast Cancer Constituents, Heterocyclic Aromatic Amines and Mycotoxins. IARC Risk by Occupation." American Journal of Industrial Medicine 30 (4): Monographs on the Evaluation of Carcinogenic Risks to Humans, vol. 430­37. 56. Lyons: IARC Press. Coogan, P. F., L. Rosenburg, S. Shapiro, and M. Hoffmann. 1999. ------. 1994. Schistosomes, Liver Flukes and Helicobacter pylori. IARC "Lactation and Breast Carcinoma in a South African Population." Monographs on the Evaluation of Carcinogenic Risks to Humans, vol. Cancer 86: 982­89. 61. Lyons: IARC Press. Cook, P. J., and D. P. Burkitt. 1971. "Cancer in Africa." British Medical ------. 1996. Human Immunodeficiency Viruses and Human T-Cell Bulletin 27: 14­20. Lymphotrophic Viruses. IARC Monographs on the Evaluation of Cook-Mozaffari, P., R. Newton, V. Beral, and D. P. Burkitt. 1998. "The Carcinogenic Risks to Humans, vol. 67. Lyons: IARC Press. Geographical Distribution of Kaposi Sarcoma and of Lymphomas in ------. 1997. Epstein-Barr Virus and Kaposi Sarcoma Herpesvirus/Human Africa before the AIDS Epidemic." British Journal of Cancer 78: 1521­28. Herpesvirus 8. IARC Monographs on the Evaluation of Carcinogenic de Thé, G., A. Geser, N. E. Day, P. M. Tukei, E. H. Williams, D. P. Beri, P. G. Risks to Humans, vol. 70. Lyons: IARC Press. Smith, et al. 1978. "Epidemiological Evidence for Causal Relationship ------. 1999. Hormonal Contraception and Post-menopausal Hormonal between Epstein Barr Virus and Burkitt's Lymphoma from Ugandan Therapy. IARC Monographs on the Evaluation of Carcinogenic Risks Prospective Study." Nature 274: 756­61. to Humans, vol. 72. Lyons: IARC Press. Dorn, H. F., and S. J. Cutler. 1959. Morbidity from Cancer in the United ------. 2002. Some Traditional Herbal Medicines, Some Mycotoxins, States: Parts I and II. Public Health Monograph 56. Washington, DC: Naphthalene and Styrene. IARC Monographs on the Evaluation of U.S. Department of Health, Education and Welfare. Carcinogenic Risks to Humans, vol. 82. Lyons: IARC Press. Dorrington, R., D. Bourne, D. Bradshaw, R. Laubscher, and I. M. Timaeus. ------. 2004. Tobacco Smoke and Involuntary Smoking. IARC Mono- 2001. The Impact of HIV/Aids on Adult Mortality. MRC Technical graphs on the Evaluation of Carcinogenic Risks to Humans, vol. 83. Report. Cape Town: MRC. http://www.mrc.ac.za/bod/bod.htm. Lyons: IARC Press. Dukers, N. H., and G. Rezza. 2003. "Human Herpesvirus 8 Epidemiology: International Collaboration on HIV and Cancer. 2002. "Review: What We Do and Do Not Know." AIDS 17: 1717­30. Breastfeeding Is Associated with Reduced Risk of Breast Cancer." Echimane, A. K., A. A. Ahnoux, I. Adoubi, S. Hien, K. M'Bra, A. Lancet 20: 187­95. D'Horpock, M. Diomande, D. Anongba, I. Mensah-Adoh, and D. M. Jaffe, E. S., N. L. Harris, H. Stein, and J. W. Vardiman, eds. 2001. WHO Parkin. 2000. "Cancer Incidence in Abidjan, Ivory Coast: First Results Classification of Tumours: Pathology and Genetics of Tumours of the from the Cancer Registry, 1995­1997." Cancer 89 (3): 653­63. Haematopoietic and Lymphoid Tissues. Lyons: IARC Press. Ferlay, J., F. Bray, P. Pisani, and D. M. Parkin. 2004. "GLOBOCAN 2002: Jaskiewicz, K., H. D. Lowrens, C. W. Woodroof, M. J. van Wyk, and S. K. Cancer Incidence, Mortality and Prevalence Worldwide." Version 2.0, Price. 1989. "The Association of Campylobacter pylori with Mucosal IARC CancerBase 5. Lyons: IARC Press. Pathological Changes in a Population at Risk of Gastric Cancer." South Flegg Mitchell, H. 1966. "Sociological Aspects of Cancer Rate Surveys in African Medical Journal 75: 417­19. Africa." Journal of the National Cancer Institute Monographs 25: 151­70. Key, T. 1995. "Risk Factors for Prostate Cancer." In Preventing Prostate Geser, A., G. Brubaker, and C. C. Draper. 1989. "Effect of a Malaria Cancer. Screening versus Chemoprevention, ed. R. T. D. Oliver, A. Suppression Program on the Incidence of African Burkitt's Belldegrun, and P. F. M. Wrigley. Cancer Surveys, vol. 23. Cold Spring Lymphoma." American Journal of Epidemiology 129: 740­52. Harbor, NY: Cold Spring Harbor Laboratory Press. Geser A., G. de Thé, G. Lenoir, N. E. Day, and E. H. Williams. 1982. "Final Kidd, M., A. J. Lastovica, J. C. Atherton, and J. A. Louw. 1999. Case Reporting from the Ugandan Prospective Study of the "Heterogeneity in the Helicobacter Pylori VacA and CagA Genes: Relationship between EBV and Burkitt's Lymphoma." International Association with Gastroduodenal Disease in South Africa?" Gut 45: Journal of Cancer 29: 397­400. 499­502. Grulich, A. E., V. Beral, and A. J. Swerdlow. 1992. "Kaposi Sarcoma in Kirk, G. D., O. A. Lesi, M. Mendy, A. O. Akano, O. Sam, J. J. Goedert, England and Wales before the AIDS Epidemic." British Journal of P. Hainaut, A. Hall, H. Whittle, and R. Montesano. 2004. "The Gambia Cancer 66: 1135­37. Liver Cancer Study. Infection with Hepatitis B and C: The Risk of Hepatocellular Carcinoma in West Africa." Hepatology 39: 211­19. Harington, J. S., E. M. Bradshaw, and N. D. McGlashan. 1983. "Changes in Primary Liver and Oesophageal Cancer Rates among Black Kogevinas, M., N. Pearce, M. Susser, and P. Boffetta, eds. 1997. Social Goldminers, 1964­1981. South African Medical Journal 64: 650. Inequalities and Cancer. IARC Scientific Publications 138. Lyons: IARC Hayes, R. B., L. M. Pottern, H. Strickler, C. Rabkin, V. Pope, G. M. Swanson, Press. R. S. Greenberg, et al. 2000. "Sexual Behaviour, STDs and Risks for Kolonel, L. N. 1996. "Nutrition and Prostate Cancer." Cancer Causes and Prostate Cancer." British Journal of Cancer 82 (3): 718­25. Control 7: 83­94. Henderson, B. E., R. Ross, and L. Bernstein. 1988. "Estrogens as a Cause of Koulibaly, M., I. S. Kabba, A. Cisse, S. B. Diallo, M. B. Diallo, N. Keita, Human Cancer: The Richard and Hinda Rosenthal Foundation Award N. D. Camara, M. S. Diallo, B. S. Sylla, and D. M. Parkin. 1997. "Cancer Lecture." Cancer Research 48: 246­53. Incidence in Conakry, Guinea: First Results from the Cancer Registry Hunter, D. J., D. Spiegelman, H.-O. Adami, L. Beeson, P. A. van den Brandt, 1992­1995." International Journal of Cancer 70 (1): 39­45. A. R. Folsom, G. E. Fraser, et al. 1996. "Cohort Studies of Fat Intake and Labadarios, D., A. R. Walker, R. Blaauw, and B. F. Walker. 1996."Traditional the Risk of Breast Cancer--A Pooled Analysis." New England Journal of Diets and Meal Patterns in South Africa." World Review of Nutrition Medicine 334: 356­61. and Diet 79: 70­108. IARC (International Agency for Research on Cancer). 1986. Tobacco Levin, C. V., B. El-Gueddari, and A. Meghzifene. 1999. "Radiation Therapy Smoking. IARC Monographs on the Evaluation of Carcinogenic Risks in Africa: Distribution and Equipment." Radiotherapy Oncology 52 (1): to Humans, vol. 38. Lyons: IARC Press. 79­84. ------. 1988. Alcohol Drinking. IARC Monographs on the Evaluation of Louw, J. A., M. S. G. Kidd, A. F. Kummer, K. Taylor, U. Kotze, and D. Carcinogenic Risks to Humans, vol. 44. Lyons: IARC Press. Hanslo. 2001. "The Relationship between Helicobacter pylori Infection, 302 | Freddy Sitas, Max Parkin, Zvavahera Chirenje, Lara Stein, Nokuzola Mqoqi, and Henry Wabinga the Virulence Genotypes of the Infecting Strain and Gastric Cancer in Oettlé, A. G., and J. Higginson. 1966. "Age Specific Cancer Incidence Rates the African Setting." Helicobacter 6: 268­73. in the South African Bantu: Johannesburg 1953­1955." South African Madhava, V., C. Burgess, and E. Drucker. 2002. "Epidemiology of Chronic Journal of Medical Science 31: 21­41. Hepatitis C Virus Infection in Sub-Saharan Africa." Lancet Infectious Pacella-Norman, R., M. I. Urban, F. Sitas, H. Carrara, R. Sur, M. Hale, Disease 2: 293­302. P. Ruff, et al. 2002. "Risk Factors for Oesophageal, Lung, Oral and Mandelblatt, J., P. Kanetsky, L. Eggert, and K. Gold. 1999. "Is HIV Infection Laryngeal Cancers in Black South Africans." British Journal of Cancer a Cofactor for Cervical Squamous Cell Neoplasia?" Cancer 86 (11): 1751­56. Epidemiology Biomarkers and Prevention 8: 97­106. Pagano, J. S., G. Jimenez, N. S. Sung, N. Raab-Traub, and J. C. Lin. 1992. Mandishona, E., A. P. MacPhail, V. R. Gordeuk, M. A. Kedda, A. C. "Epstein-Barr Viral Latency and Cell Immortalization as Targets for Paterson, T. A. Rouault, M. C. Kew. 1998. "Dietary Iron Overload as a Antisense Oligomers." Annals of the New York Academy of Sciences 660: Risk Factor for Hepatocellular Carcinoma in Black Africans." 107­16. Hepatology 27: 1563­66. Parkin, D. M., J. Ferlay, M. Hamdi-Cherif, F. Sitas, J. O. Thomas, H. Manning, E. B., J. I. Mann, E. Sophangisa, and A. S. Truswell. 1971. Wabinga, and S. L. Whelan. 2003. Cancer in Africa--Epidemiology and "Dietary Patterns in Urbanised Blacks." South African Medical Journal Prevention. IARC Scientific Publications 153. Lyons: IARC Press. 48: 488­98. Parkin, D. M., P. Pisani, A. D. Lopez, and E. Masuyer. 1994. "At Least One Mathers, C. D., D. M. Fat, M. Inoue, C. Rao, and A. D. Lopez. 2005. in Seven Cases of Cancer Is Caused by Smoking. Global Estimates for "Counting the Dead and What They Died From: An Assessment of the 1985." International Journal of Cancer 59: 494­504. Global Status of Cause of Death Data." Bulletin of the World Health Parkin, D. M., and A. J. Sasco. 1993. "Lung Cancer: Worldwide Variation in Organization 83: 171­77. Occurrence and Proportion Attributable to Tobacco Use." Lung Cancer Mbulaiteye, S. M., E. T. Katabira, H. Wabinga, D. M. Parkin, P. Virgo, 9: 1­16. R. Ochai, M. Workneh, A. Coutinho, E. A. Engels. Forthcoming. Parkin, D. M., H. R. Wabinga, S. Nambooze, and F. Wabwire-Mangen. "Spectrum of Cancers among HIV-Infected Persons in Africa: The 1999. "AIDS Related Cancers in Africa. Maturation of the Epidemic in Uganda AIDS-Cancer Registry Match Study." International Journal Uganda." AIDS 13: 2563­70. of Cancer. Peto, R., A. D. Lopez, J. Boreham, M. Thun, and C. Heath. 2004. Mortality McFarlane, G., D. Forman, F. Sitas, and G. Lachlan. 2001. "A Minimum from Smoking in Developed Countries 1950­2000. Oxford: Oxford Estimate for the Incidence of Gastric Cancer in Eastern Kenya." British University Press. Journal of Cancer 85 (9): 1322­25. Pisani, P., D. M. Parkin, N. Munoz, and J. Ferlay. 1997. "Cancer and Miller, A. B. 1992. Cervical Cancer Screening Programmes. Managerial Infection: Estimates of the Attributable Fraction in 1990." Cancer Guidelines. Geneva: WHO. Epidemiology, Biomarkers Prevention 6: 387­400. Motala, A. A. 2002. "Diabetes Trends in Africa." Diabetes Metabolism Prates, M. D., and F. O. Torres. 1965. "A Cancer Survey in Lorenco Research Reviews 18 (Suppl. 3): S14­20. Marques, Portuguese East Africa." Journal of the National Cancer Muir, C., J. Waterhouse, T. Mack, J. Powell, S. Whelan, M. Smans, and F. Institute 35: 729­57. Casset. 1987. Cancer Incidence in Five Continents. Vol. 5. IARC Prentice, R. L., and L. Sheppard. 1990. "Dietary Fat and Cancer: Scientific Publication 88. Lyons: IARC Press. Consistency of Epidemiologic Data and Disease Prevention that May Muir-Grieve, J. "South Africa, Cape Province." 1960. In Cancer Incidence in Follow from a Practical Reduction in Fat Consumption." Cancer Five Continents, vol. 2, ed. R. Doll, C. Muir, and J. Waterhouse, 98­109. Causes and Control 1: 81­97. Berlin: Springer Verlag. Rabkin, C. S., R. J. Biggar, and J. W. Horm. 1991. "Increasing Incidence of Mzileni, O., F. Sitas, K. Steyn, H. Carrara, and P. Bekker. 1999. "Lung Cancers Associated with the Human Immunodeficiency Virus Cancer, Tobacco and Environmental Factors in the African Population Epidemic." International Journal of Cancer 47: 692­96. of the Northern Province, South Africa." Tobacco Control 8: 398­401. Reeves, G. 1996. "Breast Cancer and Oral Contraceptives--The Evidence Newton, R., A. Grulich, V. Beral, B. Sindikubwabo, P.-J. Ngilimana, A. So Far." Cancer Causes and Control 7: 495­96. Nganyira, and D. M. Parkin. 1995. "Cancer and HIV Infection in Remick, S. C. 1995. "Acquired Immunodeficiency Syndrome-Related Non- Rwanda." Lancet 345 (8961): 1378­79. Hodgkin Lymphoma." Cancer Control 2: 97­103. Newton, R., P.-J. Ngilimana, A. Grulich, V. Beral, B. Sindikubwabo, A. Reza, A., J. A. Mercy, and E. Krug. 2001. "Epidemiology of Violent Deaths Nganyira, and D. M. Parkin. 1996. "Cancer in Rwanda." International in the World." Injury and Prevention 7: 104­11. Journal of Cancer 66: 75­81. Robertson, M. A. 1969. "Clinical Observations on Cancer Patterns at the Newton, R., F. Sitas, M. Dedicoat, and J. L. Ziegler. 2002. "HIV Infection Non-White Hospital Baragwanath, Johannesburg, 1948­1964." South and Cancer." In AIDS in Africa, 2nd ed., ed. M. Essex, S. Mboup, P. J. African Medical Journal 26: 915­31. Kanki, R. G. Marlink, and S. D. Tlou. New York: Kluwer Academic. Rosenberg, L., J. P. Kelly, S. Shapiro, M. Hoffman, and D. Cooper. 2002. Newton, R., J. Ziegler, V. Beral, and the Uganda Kaposi Sarcoma Study "Risk Factors for Breast Cancer in South African Women." South Group. 2001. "A Case-Control Study of Human Immunodeficiency African Medical Journal 92: 447­48. Virus Infection and Cancer in Adults and Children Residing in Ross, R. K., M. C. Pike, G. A. Coetzee, J. K. Reichardt, M. C. Yu, H. Kampala, Uganda." International Journal of Cancer 92: 622­27. Feigelson, F. Z. Stanczyk, L. N. Kolonel, and B. E. Henderson. 1998. O'Conor, G. T. 1970. "Persistent Immunologic Stimulation as a Factor in "Androgen Metabolism and Prostate Cancer: Establishing a Model of Oncogenesis with Special Reference to Burkitt's Tumour." Annual Genetic Susceptibility." Cancer Research 58: 4497­4504. Journal of Medicine 48: 279­85. Sammon, A. M. 1992. "A Case-Control Study of Diet and Social Factors in Oettlé, A. G. 1962. "Geographical and Racial Differences in the Frequency Cancer of the Oesophagus in Transkei." Cancer 69: 860­65. of Kaposi Sarcoma as Evidence of Environmental or Genetic Causes." Sasco, A., A. B. Lowels, and P. Pasker de Jong. 1993. "Epidemiology of Male Acta Unio Internationalis Contra Cancrum 18: 330­63. Breast Cancer. A Meta-Analysis of Published Cases-Control Studies ------. 1964. "Cancer in Africa Especially in Region South of the Sahara." and Discussion of Selected Etiological Factors." International Journal of Journal of the National Cancer Institute 33: 383­439. Cancer 53: 538­49. Cancers | 303 Sathar, M. A., A. E. Simjee, D. F. Wittenberg, and A. M. Mayat. 1994. Infection, HIV Infection and Cervical Cancer in Tanzania, East Africa." "Seroprevalence of Helicobacter pylori Infection in Natal/KwaZulu, International Journal of Cancer 51: 515­21. South Africa." European Journal of Gastroenterology and Hepatology 6: Timaeus, I. M. 1999. "Mortality in Sub-Saharan Africa." In Health and 37­41. Mortality: Issues of Global Concern, ed. J. Chamie and R. L. Cliquet, Schonland, M., and E. Bradshaw. 1968. "Cancer in the Natal African and 108­31. New York: United Nations Population Division. Indian, 1964­1966." International Journal of Cancer 3: 304­16. Turner, P. C., A. Sylla, Y. Y. Gong, M. S. Diallo, A. E. Sutcliffe, A. J. Hall, and Schultz, T. F., C. H. Boshoff, and R. A. Weiss. 1996. "HIV Infection and C. P. Wild. 2005. "Reduction in Exposure to Carcinogenic Aflatoxins by Neoplasia." Lancet 348: 587­91. Postharvest Intervention Measures in West Africa: A Community- Seedat, Y. K. 2000. "Hypertension in Developing Nations in Sub-Saharan Based Intervention Study." Lancet 365: 1950­56. Africa." Journal of Human Hypertension 14: 739­47. University of Zimbabwe/JHPIEGO Cervical Cancer Project. 1999. "Visual Shapiro,S.,L.Rosenberg,M.Hoffman,H.Truter,D.Cooper,S.Rao,D.Dent, Inspection with Acetic Acid for Cervical Cancer Screening: Test et al. 2000. "Risk of Breast Cancer in Relation to the Use of Injectable Qualities in a Primary-Care Setting." Lancet 353 (9156): 869­73. Progestogen Contraceptives and Combined Estrogen/Progestogen Van Rensburg, S. J. 1981. "Epidemiologic and Dietary Evidence for a Contraceptives." American Journal of Epidemiology 151: 396­403. Specific Nutritional Predisposition to Esophageal Cancer." Journal of Erratum in American Journal of Epidemiology 151: 1134. the National Cancer Institute 67: 243­51. Shibata, A., and A. S. Whittemore. 1997. "Genetic Predisposition to Wabinga, H. R., D. M. Parkin, F. Wabwire-Mangen, and J. W. Mugerwa. Prostate Cancer: Possible Explanations for Ethnic Differences in Risk." 1993. "Cancer in Kampala, Uganda, in 1989­91: Changes in Incidence Prostate 32: 65­72. in the Era of AIDS." International Journal of Cancer 54: 26­36. Sitas, F., W. R. Bezwoda, V. Levin, P. Ruff, M. C. Kew, M. J. Hale, H. Carrara, Wabinga, H. R., D. M. Parkin, F. Wabwire-Mangen, and S. Nambooze. et al. 1997. "Association between Human Immunodeficiency Virus 2000. "Trends in Cancer Incidence in Kyadondo County, Uganda, Type 1 Infection and Cancer in the Black Population of Johannesburg 1960­1997." British Journal of Cancer 82: 1585­92. and Soweto, South Africa." British Journal of Cancer 75: 1704­7. Walker, A. R. P., B. F. Walker, S. Funani, and A. J. Walker. 1989. Sitas, F., H. Carrara, V. Beral, R. Newton, G. Reeves, D. Bull, U. Jentsch, "Characteristics of Black Women with Breast Cancer in Soweto, South et al. 1999. "The Seroepidemiology of HHV-8/KSHV in a Large Africa." Cancer Journal 2: 316­19. Population of Black Cancer Patients in South Africa." New England Journal of Medicine 340: 1863­71. Walker, A. R. P., B. F. Walker, N. G. Tsotetsi, C. Sebitso, D. Siwedi, and A. J. Walker. 1992. "Case-Control Study of Prostate Cancer in Black Patients Sitas, F., J. Madhoo, and J. Wessie. 1998. "Incidence of Histologically in Soweto, South Africa." British Journal of Cancer 65: 438­41. Diagnosed Cancer in South Africa 1993­1995." National Cancer Registry, South African Institute for Medical Research, Johannesburg. Walker, R. W., D. G. McLarty, H. M. Kitange, D. Whiting, G. Masuki, D. M. Mtasiwa, H. Machibya, N. Unwin, and K. G. Alberti. 2000. "Stroke Sitas, F., R. Pacella-Norman, H. Carrara, M. Patel, P. Ruff, R. Sur, U. Jentsch, Mortality in Urban and Rural Tanzania." Adult Morbidity and et al. 2000. "The Spectrum of HIV-1 Related Cancers in South Africa." Mortality Project. Lancet 355: 1684­87. International Journal of Cancer 88: 489­92. Whittle, H. C., J. Brown, K. Marsh, M. Blackman, O. Jobe, and F. Shenton. Sitas, F., M. Urban, D. Bradshaw, D. Kielkowski, S. Bah, and R. Peto. 2004. 1990. "The Effects of Plasmodium falciparum Malaria on Immune "Tobacco Attributable Deaths in South Africa." Tobacco Control 13: Control of B Lymphocytes in Gambian Children." Clinical and 396­99. Experimental Immunology 80: 213­18. Skinner, M. E. G., D. M. Parkin, A. P. Vizcaino, and A. Ndhlovu. 1993. WHO (World Health Organization). 1997. Tobacco or Health: A Global Cancer in the African Population of Bulawayo, Zimbabwe, 1963­1977. Status Report. Geneva: WHO. IARC Technical Report 15. Lyons: IARC Press. ------. 2004. The World Health Report 2004--Changing History. Geneva: Somdyala, N. I., W. F. Marasas, F. S. Venter, H. F. Vismer, W. C. Gelderblom, WHO. and S. A. Swanevelder. 2003. "Cancer Patterns in Four Districts of the Transkei Region--1991­1995." South African Medical Journal 93 (2): Wild, C. P., and A. J. Hall. 2000. "Primary Prevention of Hepatocellular 144­48. Carcinoma in Developing Countries." Mutation Research 462: 381­93. Ssali, J. C., A. Gakwaya, and E. Katangole-Mbidde. 1995. "Risk Factors for Williams, E. H., P. G. Smith, N. E. Day, A. Geser, J. Ellis, and P. Tukei. 1978. Breast Cancer in Ugandan Women: A Case Control Study." Eastern and "Space-Time Clustering of Burkitt's Lymphoma in West Nile District Central African Journal of Surgery 1: 9­13. of Uganda: 1961­1975." British Journal of Cancer 37: 805­9. ter Meulen, J., H. C. Eberhardt, J. Luande, H. N. Mgaya, J. Chang-Claude, Wright, D. H. 1973. "Lympho-Reticular Neoplasms." Recent Results in H. Mtiro, M. Mhina, et al. 1992. "Human Papillomavirus (HPV) Cancer Research 41: 270­91. 304 | Freddy Sitas, Max Parkin, Zvavahera Chirenje, Lara Stein, Nokuzola Mqoqi, and Henry Wabinga Chapter 21 Cardiovascular Disease Anthony Mbewu and Jean-Claude Mbanya The burden of cardiovascular disease (CVD) in the world is be middle-aged people. The poor will suffer dispropor- enormous and growing, and the majority of those affected tionately as a consequence of their higher disease risk and are in developing countries (Beaglehole and Yach 2003; limited access to health care. The financial and social costs Mbewu 1998). In 2002 it was estimated that 29 percent of of this CVD epidemic are likely to have a negative impact deaths worldwide (16.7 million deaths) were due to CVD on development and the alleviation of poverty (http:// and that 43 percent of global morbidity and mortality, www.ichealth.org). measured in disability-adjusted life years (DALYs), was African countries therefore face a double burden as they caused by CVD (WHO 2002). Furthermore, 78 percent of struggle to cope with the burden of communicable diseases global mortality and 86 percent of mortality and morbidity and diseases associated with lack of socioeconomic from CVD occurs in developing countries. By 2020 it is development--the "unfinished agenda." Furthermore, their estimated that CVD will become the leading cause of the predicament is only likely to worsen, because the majority of global health burden, accounting for 73 percent of total their populations are under 35 years of age, and the deter- global mortality and 56 percent of total morbidity (Murray minants and risk factors for CVD are already prevalent and and Lopez 1996; Reddy and Yusuf 1998). increasing within this age group. Africa has not been spared this global tide of CVD. In The relative cost of the epidemic of CVD is likely to be most African countries CVD is now the second most com- higher than in upper-income countries, where CVD prima- mon cause of death after infectious disease, accounting rily affects the elderly. In African countries more than half of for 11 percent of total deaths (WHO 1999); and CVD is a CVD deaths occur among people between 30 and 69 years major cause of chronic illness and disability. Projections of age, an age 10 years or more below the equivalent group from the Global Burden of Disease Project suggest that from in Europe and North America (http://www.ichealth.org). In 1990 to 2020, the burden of CVD faced by African countries Ghana, for example, where cerebral hemorrhage is a leading will double. A large proportion of the victims of CVD will cause of death, the average age at which people die from this 305 cause is 55 years (http://www.ichealth.org). Death and increasing fat and calorie consumption, and decreasing disability in middle age have major social and economic exercise. Longer periods of exposure to these determi- consequences, depriving families of parents, workplaces of nants because of longer life expectancy have increased the employees, and communities of leaders. Patients denied rates of chronic disease. access to health care for CVD or deterred by high costs from seeking it will cause the public health systems to incur even Moreover, whereas European and North American pop- greater health care costs in the long run as a result of the ulations experienced similar changes in demography, deter- need to treat the same patients later at greater expense minants, and disease rates over the course of a few centuries, because the disease is more advanced. African countries are passing through similar transitions The potential costs of this CVD epidemic for African in just a few decades (http://www.ichealth.org). This forced countries are staggering. Cardiovascular disease (direct and pace of globalization has resulted in the "export of risk indirect) is estimated to cost the United States about factors" from the West such as tobacco, refined foods, and US$300 billion annually, equal to the entire gross domestic lifestyles with high CVD risk that are portrayed on televi- product of the African continent. Clearly, even a fraction of sion and film (Mbewu 1998). such cost has the potential to cause enormous damage to the Still, the process of epidemiological transition seems to economies and development trajectories of African coun- be different in Africa than in developed and other developing tries. In this way, the growing CVD epidemic in Africa will countries, where it is mainly marked by the explosion of increase already unacceptable levels of inequity in access to coronary heart disease (CHD). Indeed, although the health care services. epidemic of CHD was heralded in the 1980s (Ogunnowo, The overall health of African nations will not improve, Odesanmi, and Andy 1986), in Africa it is still awaited, nor will their level of development, unless they deal with this although hemorrhagic stroke is already a leading cause of epidemic of CVD. Furthermore, in an increasingly integrated mortality and morbidity (Walker et al. 2003). Furthermore, global economy the CVD epidemic in developing countries dilated cardiomyopathy is particularly prevalent in Sub- will divert economic goods to CVD care, resulting in a Saharan Africa, presumably owing to nutritional and viral reversal of developmental efforts; productivity will decline factors. The epidemiological transition was readily apparent because of the loss of more productive citizens; and in the changes in causes of mortality in the Seychelles over consumer markets will shrink as a result of loss of the the past 30 years (table 21.1) (Bovet 1995). purchasing power of these citizens. More reasons to suspect an impending epidemic of CVD in Sub-Saharan Africa include the recent finding that poor socioeconomic conditions in childhood determine CVD in middle age as strongly as do CVD risk factors in middle age THE EPIDEMIOLOGICAL TRANSITION in the same individuals (Lawlor, Smith, and Ebrahim 2002). The process responsible for these global shifts in CVD Furthermore, according to the Barker Hypothesis, poor fetal mortality is termed the "epidemiological transition" growth has been shown to be associated with hypertension, (Omran 1971). Three main drivers fuel this transition: Table 21.1 The Epidemiological Transition in the Seychelles, 1976 and 1994 · Declining infant and child mortality has led to rapid (percent) demographic changes resulting in large increases in the number of individuals surviving until middle and older Causes of death 1976 1994 age, when chronic diseases become manifest--the so- Infectious and parasitic 9.9 5.5 called demographic transition. By 2025 it is estimated Total CVD 26.3 39.5 that the number of Africans over 60 years old will Circulatory system 12.6 16.4 increase from 39 million to 80 million. Hypertensive disease -- 8.4 · Falling death rates from communicable diseases have Ischemic heart disease -- 8.0 accompanied socioeconomic development and improved Cerebrovascular disease 13.7 5.2 vaccination and other primary health care services. Cancer 8.2 16.2 · Changes have occurred in environmental and behavioral Source: Bovet 1995. determinants of CVD, such as increasing tobacco use, Note: -- not available. 306 | Anthony Mbewu and Jean-Claude Mbanya and CVD, in later life. However, a study carried out in Verbal autopsy has been shown to be an economical and Nigeria failed to demonstrate this effect (Law et al. 2001). useful way of improving the quality of cause-of-death infor- The current impoverishment of much of Sub-Saharan mation when health workers have minimal training. For Africa may paradoxically result in an epidemic of CVD in example, in an investigation of causes of death in women of middle age for those who survive the ravages of poverty- childbearing age in Guinea-Bissau, 70 percent of deaths associated communicable diseases, such as AIDS, tubercu- could be attributed to a specific disease or condition (Sen losis, pneumonia, and malaria. Because fetal growth retar- and Bonita 2000). dation is associated with chronic undernutrition among women, improvement in the nutrition and health of girls Cardiovascular Surveillance Systems and young women may be important in preventing CVD in developing countries. Data on the incidence and prevalence of CVD is scanty. Surveillance should become a critical component in the strategies adopted in Sub-Saharan Africa to deal with the SOURCES OF DATA burgeoning epidemic of CVD. Sentinel surveillance is likely to be the preferred methodology, whereby monitoring of Sources of data on CVD rates in Sub-Saharan Africa are disease episodes is periodically conducted at sentinel sites generally lacking, and when present, are often of poor quality. that are broadly representative of the general population. Much of the available data comes from individual studies, This methodology is preferred because Sub-Saharan coun- often hospital-based, with small numbers of participants. tries have limited resources, and their health systems cannot Often the different data sources are heterogenous in monitor every single disease episode. The systems set in methodology and cannot be compared; and systematic and place should include behavioral surveillance, often dubbed regular surveillance systems are almost totally absent, "second generation surveillance." Demographic health making it difficult to plot changes in CVD rates over time. surveys are cheaper to administer but suffer from the pitfall Nevertheless, the available data sources do give some idea of of their cross-sectional design. Surveillance will not only the nature and magnitude of CVD in Sub-Saharan Africa monitor the prevalence of CVD but will also help gauge the and of the changes in the nature and rates of CVD that have impact of primary prevention strategies. taken place over the past 50 years. Crude Mortality versus Age-Standardized Mortality Measures of Cardiovascular Mortality The importance of CVD mortality in Sub-Saharan Africa Globally, cause-of-death data based on the death certificate countries tends to be underestimated, because crude as provided to the data bank of the World Health mortality rates rather than age-standardized mortality rates Organization (WHO) is available in only 77 countries. are used. Thus the appalling figures for infant mortality There is wide regional variation in coverage by national vital from infectious disease and diseases of poverty tend to over- registration systems, ranging from 80 percent population shadow all other causes of death. Yet, from the perspective of coverage in the European region to less than 5 percent in the citizens of Sub-Saharan countries when contemplating their Eastern Mediterranean and African regions of WHO (Sen own mortality, or the policy maker when considering causes and Bonita 2000). The most serious gap is for adult mor- of death among economically active people, age-specific tality, crucial if one is to gauge the true extent of CVD in the mortality ratios are critical. They are also important for the developing world and monitor trends over time. planner predicting the patterns of death in years to come. The 10-fold variation in infant mortality between different Age-standardized mortality ratios make it clear that the regions of the world is largely due to communicable disease, long-heralded epidemic of CVD in Africa has, at least in malnutrition, and poverty, whereas the cause of death in some African countries, already arrived. Indeed, in some adults age 15 to 60 is almost entirely due to noncommuni- middle-income countries, such as South Africa, age- cable diseases (NCDs) and injury (Murray and Lopez 1997). standardized mortality rates are higher than for Scotland Men in Sub-Saharan Africa are three times more likely to and Finland, and they are exceeded only by rates in the die prematurely than men in Western industrial populations former socialist economies of Europe (figure 21.1) as a result of AIDS and violence. (Bradshaw et al. 2003). Cardiovascular Disease | 307 Figure 21.1 Cardiovascular Disease, Age-Standardized Rates was common among young people in Kampala, Uganda, in the World, 1994­2000 where autopsy data demonstrated cardiac disease to be a (per 100,000 people) cause of death in 15.3 percent of males older than 60 years Bulgaria 1998 and in 13.1 percent of males younger than 30 years (Drury Kazakhstan 1997 Moldova 1996 1972). In 1980 the Zambian Ministry of Health report of Ukraine 1998 Russian Federation 1997 admissions and causes of death in all government, mining, Romania 1998 Belarus 1998 and mission hospitals listed 2.06 percent of deaths in those Latvia 1998 Kyrgyz Republic 1998 Mauritius 1997 under age 14 years as cardiac in origin compared with Azerbaijan 1997 Estonia 1998 14.50 percent in those over age 14. As regards morbidity, The FYR of Macedonia 1997 Hungary 1998 0.62 percent of hospital admissions of children under 14 Slovak Republic 1995 Croatia 1997 in that same study were cardiac in origin compared with Lithuania 1997 Poland 1996 4.23 percent of those individuals over age 14 (Hutt 1990). Armenia 1997 Czech Republic 1998 South Africa 2000 By 1990, CVD had become the third most common cause Argentina 1996 Brazil, South, SE & Mid-West 1995 of death in a prospective autopsy study in 90 of the 167 Ireland 1996 UK: Scotland 1997 deaths in one year at Tshepong Hospital in the North West Austria 1998 Slovenia 1997 Province of South Africa (Steenkamp, Simson, and Theron Malta 1997 Portugal 1998 1992). In patients over age 35, CVD was the most common UK: Northern Ireland 1997 Finland 1996 Greece 1997 cause of death. Germany 1997 Costa Rica 1995 Cerebrovascular disease accounted for 32 percent of the New Zealand 1996 Norway 1995 CVD deaths overall. Among these, intracerebral hemor- United Kingdom 1997 United States 1997 rhage was found in 50 percent and cerebral infarction in UK: England and Wales 1997 Denmark 1996 29 percent of cases. Fifty-seven percent of cardiovascular Sweden 1996 Iceland 1995 Luxembourg 1997 deaths were due to cardiac conditions, the most common Italy 1995 Australia 1995 being pulmonary hypertension (31 percent), dilated car- Belgium 1994 Netherlands 1997 diomyopathy and chronic rheumatic valvular disease Israel 1996 Spain 1995 (17 percent each), and hypertensive heart disease (14 per- Canada 1997 Cayman Islands 1994 cent). Only 3 percent of the examined vessels had signs of France 1996 Japan 1997 severe atherosclerosis. The clinical diagnosis was the same as 0 100 200 300 400 500 600 700 the final autopsy diagnosis in only 38 percent of cases, Source: Debbie Bradshaw (personal communication) and WHO Mortality Database, WHO Statistical Information System at http://www3.who.int/whosis/ (accessed June 2005). emphasizing the importance of performing autopsies to obtain reliable mortality statistics in African countries. In a prospective study among elderly patients in Kenyatta THE CHANGING PREVALENCE, INCIDENCE, AND National Hospital, Nairobi, Kenya, in 1991 to 1992, clinical PATTERN OF CARDIOVASCULAR DISEASE evidence of CVD was present in 40 percent of the patients evaluated; 54 percent were hypertensive, 53 percent had The relative and absolute importance of CVD in countries arrhythmia, and 49 percent had congestive cardiac failure of Sub-Saharan Africa is thought to be increasing, but infor- (Lodenyo, McLigeyo, and Ogola 1997). A prospective study mation on morbidity, mortality, and prevalence of disease of 708 subjects with CVD was conducted between January and risk factors is scanty (Muna 1993b; Razum 1996). Data 1992 and December 1995 in Ghana (Amoah 2000). from 20 years ago consisted largely of hospital records, Participants were evaluated clinically, with ancillary labora- which reported that 8 to 12 percent of hospital admissions tory tests, chest X-ray, electrocardiography, and two- were due to heart disease in several countries in southern dimensional echocardiography with doppler and color flow Africa (Vaughan 1977); hypertensive and rheumatic heart mapping. Hypertensive heart disease (n 133), RHD (n disease (RHD) accounted for 40 to 60 percent of these 123), idiopathic cardiomyopathy (n 103), congenital admissions. There were few recorded episodes of ischemic heart disease (n 90), and coronary artery disease (CAD; cardiac events. By 1977 such causes of death as syphilitic n 80) were the major causes of cardiovascular morbidity. heart disease were rarely recorded, despite the continuing The mean age of the subjects was 41.6; peak incidence of high frequency of acute syphilis. In the 1970s cardiac disease CVD occurred during the decile 40­49 years of age. 308 | Anthony Mbewu and Jean-Claude Mbanya A retrospective study in 1995 in Cameroon of 312 adult Cerebrovascular Accidents patients with CVD, average age 44 years, revealed high blood The prevalence and incidence of stroke in Sub-Saharan pressure (38.5 percent), rheumatic valvular heart diseases Africa have increased over the last half century, due princi- (25.6 percent), cardiomyopathies (22.5 percent), and other pally to increased life expectancy and changes in environ- cardiovascular diseases (13.5 percent) (Kotto and Bouelet mental determinants and risk factors. The majority of 2000). Rheumatic valvulopathies were predominant among cerebrovascular accidents (CVAs) occur in young and the age group 20 to 39 years, hypertension was predominant middle-aged people and are related to hypertension. from the age of 40 years, and cardiomyopathies were Hypertension is highly prevalent in Sub-Saharan Africa and observed in the age range 20 to 60 years. is often undetected or poorly controlled. This may be the A Cameroonian study between 1992 and 1997 ranked explanation for the high proportion of hemorrhagic CVAs, coronary artery disease eighth among the CVDs registered whereas in developed countries most CVAs occur in older with a prevalence of 1.53 percent (2.42 percent in males people and are thrombotic in etiology. This has been and 0.45 percent in females). Myocardial infarction was the confirmed by clinical, radiological, and postmortem diag- most frequent clinical form of CAD observed (43 percent), nostic methods. Overall, CVAs account for 7 percent of followed by angina pectoris (23 percent), unstable angina deaths in South Africa (Statistics South Africa 1996). (20 percent), and other forms of ischemic heart disease Cross-sectional, hospital-based studies of the prevalence (13 percent). The cardiovascular risk factors were obesity and incidence of CHD and stroke and associated risk factors (80 percent), hypertension (60 percent), dyslipidemia have been carried out in South Africa, central Africa, West (43 percent), smoking (36 percent), diabetes/hyperglycemia Africa, and North Africa (Ezenwaka et al. 1997; Vorster 2002; (26 percent), and hyperuricemia (20 percent). Seventy-six Walker and Sareli 1997; Wiredu and Nyame 2001). In the percent of the patients had at least three cardiovascular risk city of Tunis in Tunisia the crude annual incidence rate of factors (Mbanya et al. 1998). stroke has been estimated at 54 per 100,000 and the preva- In multiethnic South Africa, CHD is the major cause of lence rate at 600 to 1,400 per 100,000. The incidence rate death among white people and South Africans of Indian adjusted to population at risk (greater than or equal to descent, with incidence rates of 165.3 and 101.2 per 100,000 45 years old), is about 192 per 100,000. A door-to-door people, respectively, but only 55.1 per 100,000 among survey conducted in the town of Kelibia in Tunisia showed people of mixed descent and 5.3 per 100,000 among black a prevalence rate of 720 per 100,000 when adjusted to pop- African people. Cerebrovascular disease is the most com- ulation at risk. The crude incidence rate of stroke was esti- mon cause of CVD death among those of mixed descent, mated to be between 1 and 30 per 100,000; and the stan- followed by white people and South Africans of Indian dardized rate was 68 per 100,000. Fifty percent of the stroke descent, and then black African people (73.6, 62.5, and 36.5 victims were below the age of 54 years; and one-third of per 100,000, respectively) (Bradshaw et al. 2003). them died within one week of the stroke. Overall, the age- In South Africa 90 percent of deaths are certified; of these specific rates for both sexes rose with age, with the rates for certifications over 70 percent are by a medical doctor, and women being higher at all age strata except for the group age cause-specific data are available for most deaths. A gradually 45 to 54 years (Mirabet 1990). In a South African study of shrinking proportion, currently 13 percent, are categorized stroke patients in 1998, only 20 percent of the total group as "ill-defined." It is quite possible that many deaths from understood that hypertension had probably caused their CVD could masquerade as ill-defined deaths, particularly in stroke, although 76 percent of the older group and 56 per- a country where there is still a reluctance to diagnose CVD cent of the younger group had been told at some stage that death in black people. they were hypertensive (Hale, Fritz, and Eales 1998). In a study of CVA in 21 centers in Africa, Asia, Europe, and Latin America, using computed tomography (CT) scan, magnetic resonance imaging (MRI), or cerebral angiogra- EPIDEMIOLOGY OF THE VARIOUS CVDS phy, the overall odds ratio of ischemic stroke was 2.99 (95 percent CI, 1.65­5.40) in Europe and 2.93 (2.15­4.00) in With such inadequate data sources, it is inevitable that the the non-European (developing) countries. epidemiology of CVD in Sub-Saharan Africa will be poorly Table 21.2 shows data on incidence of stroke in Sub- understood. Saharan Africa from a review of literature on hospital studies Cardiovascular Disease | 309 Table 21.2 Incidence of Stroke has been increasing since the 1980s (Hutt 1990), however, with reports of clinical IHD and increasing CHD prevalence Author and date Site Incidence (Bertrand 1992; Hutt 1990). The risk factors seem to be the Rosman 1986 Hospital based (review of literature) 101/100,000 same as in Europe, but the risk index is 2.1 to 2.7 compared Osuntokun et al. 1987 Population 58/100,000 with 3.6 in France. Myocardial infarction at 49 percent was Matenga 1997 Hospital based 30.7/100,000 the most common manifestation of CHD, followed by Source: Compiled by authors. angina pectoris at 32 percent; ischemic cardiomyopathy, 7 percent; and ventricular aneurysm, 7 percent (Bertrand 1992). Myocardial infarction in black Africans under age 40 years shows characteristics similar to those seen in in Africa (Rosman 1986), a population-based study in patients under age 40 in the West, particularly regarding the Nigeria (Osuntokun et al. 1987), and a hospital study in frequency of myocardial infarction as the first manifestation Zimbabwe (Matenga 1997). of the disease, low prevalence of coronary artery stenosis, In Mauritius between 1990 and 1994 (Sarti et al. 2000) and a relatively common finding of normal coronary arteri- the age-standardized stroke mortality for women and men ography (Nethononda et al. 2004). age 35 to 74 years was 268 per 100,000 and 138 per 100,000, The increase in CHD in Sub-Saharan Africa since the respectively. 1980s is presumably because of the increasing prevalence In Tanzania, recent verbal autopsy data demonstrated among African populations of the classical risk factors for that the age-adjusted stroke mortality rates were high CAD: smoking, a diet high in saturated fat, hypertension, (Walker et al. 2000). During the three-year observation obesity, diabetes mellitus, and lack of physical exercise. In period 11,975 deaths were recorded in three surveillance addition, life expectancy in Sub-Saharan Africa has risen areas, of which 7,629 (64 percent) were of adults age 15 years since the 1950s, meaning that more people are exposed to or older; of these, 4,088 (54 percent) were of men and 3,541 these risk factors for long enough periods to cause CAD. A (46 percent) were of women.CVD accounted for 421 (5.5 per- study of black African patients admitted to a coronary care cent) of the deaths; of these, 225 (53 percent) were of men unit with acute myocardial infarction between 1995 and and 196 (47 percent) were of women. The yearly age-adjust- 1996 showed high rates of smoking and hypertension ed rates per 100,000 in the 15-to-64-year age group for the among the patients compared with controls matched by age three project areas (urban, fairly prosperous rural, and poor and sex (Mayosi et al. 1997). rural) were 65 (95 percent CI, 39­90), 44 (31­56), and 35 A coronary angiographic study of black African patients (22­48), respectively, for men, and 88 (48­128), 33 (22­43), with acute myocardial infarction and acute ischemic syn- and 27 (16­38) for women. In a hospital-based study of 116 dromes admitted to Chris Hani Baragwanath Hospital in patients in Pretoria, one-month mortality was 33.6 percent Johannesburg, South Africa, showed a clear increase in (Rosman 1986). prevalence of risk factors for CHD in these patients com- pared with age- and sex-matched controls (Nethononda et al. 2004). In coronary angiographic studies of black Coronary Heart Disease Africans following myocardial infarction, mild to moderate CHD, clinically manifested as ischemic heart disease (IHD), coronary artery disease is often found rather than the mod- was formerly rare in Sub-Saharan Africa, again probably erate to severe artery disease found in their white and Indian largely because the majority of Africans did not live long counterparts. Indeed, CHD is the most common cause of enough to suffer the clinical manifestations of angina, acute morbidity and mortality in South Africans of Indian descent ischemic syndromes, myocardial infarction, and heart fail- (Seedat 1998). ure that usually develop in middle and old age. Still, even in The risk factor profile, then, for CHD is the same in Sub- those Africans who did live long enough for the cumulative Saharan Africa countries as in Western countries, but the effects of risk factors for CHD to take effect, CHD was rare hemoglobin S or C trait could be a risk factor for CHD up until the mid-twentieth century as evidenced by 3,500 unique to Sub-Saharan Africa. The long-term outcome of postmortem studies in Ghana (Edington 1954) in which infarction is severe and influenced by myocardial sequelae only three cases of CHD were found; of 635 cases of cardiac of imprecise origin, delayed hospitalization, absence of death in Uganda in 1966, 10 years later, less than 1 percent thrombolysis and angioplasty, and socioeconomic and of CHD was found at autopsy (Hutt and Coles 1969). CHD literacy problems. 310 | Anthony Mbewu and Jean-Claude Mbanya Mortality from CHD is much more difficult to estimate ventricular septal defect and pulmonary stenosis, giving a than that from stroke without a population-based study in prevalence of 1.8 per 1,000. Sub-Saharan Africa. Steinberg, Balfe, and Kustner (1988) RHD is a disease of poverty, related to overcrowding, reported a 25 percent decline in age-adjusted mortality from poor housing, and undernutrition and requires a multisec- IHD in South Africa, from 162 per 100,000 people in 1978 toral response for prevention and cure. It is caused by group to 121 per 100,000 in 1985. Bertrand (1992) reported an in- A beta-hemolytic streptococci. The principal methods of hospital mortality after myocardial infarction of 15 percent. control are primary and secondary prevention of strepto- A recent case-control study of 98 black South Africans coccal infection. Specifically, these preventive measures with CHD over 15 years culminated in 58 deaths from entail prompt treatment of streptococcal throat infections myocardial infarction with postmortem data available. with penicillin in primary prevention and penicillin pro- Logistic regression analysis revealed that the classical risk phylaxis following rheumatic fever in order to prevent rheu- factors in this cohort of patients operated in the same way as matic heart disease in secondary prevention. The lack of in Western populations (K. Steyn, unpublished observa- these preventive measures explains the persistence of rheu- tions). Similar results have been found in the soon-to-be matic fever and RHD in Sub-Saharan Africa, which, com- published Interheart Study of risk factors for CHD among pared with the rest of the world, has remained poor. 15,152 patients around the world compared with age- and Countries in Sub-Saharan Africa also lack adequate health sex-matched controls. This included several hundred systems for managing rheumatic fever and RHD. patients in South Africa from all ethnic groups and demon- strated that nine easily measured risk factors are associated Heart Failure with more than 90 percent of the risk. These results are consistent across all geographic regions and ethnic groups Systemic hypertension is the most common cause of heart of the world, men and women, and young and old (Yusuf failure among black Africans. In a study of 52 Gambians and et al. 2004). 55 Nigerians between ages 16 and 69 years with hypertensive heart failure, the mean duration of diagnosis of systemic hypertension among the previously known hypertensives was 4.3 years (Isezuo et al. 2000). The overall one-year sur- Rheumatic Heart Disease vival rate was 71 percent, although it was unclear whether Twenty years ago rheumatic heart disease (RHD) was the this was largely systolic or diastolic heart failure and whether most common form of cardiac disease in Sub-Saharan the cases were primarily essential hypertension or included Africa and still remains prevalent, with many young people large numbers with secondary hypertension. The prognosis in their teens and early twenties presenting with severe RHD of hypertensive heart failure among this population is poor, (Ekra and Bertrand 1992). In the 1980s RHD accounted with the first three months from onset of heart failure being for 10 to 35 percent of hospital cardiac patients in Sub- critical for survival. Early detection and control of systemic Saharan Africa (Hutt 1990) and up to 20 percent of cardiac hypertension should be more aggressively pursued. deaths noted at autopsy. Most of the cases occurred in young A 1993 study of patients admitted to Kenyatta National people. In Soweto, Johannesburg, South Africa, the inci- Hospital with congestive heart failure revealed that almost dence of RHD among primary schoolchildren was 6.9 per 32 percent had RHD, 25 percent had cardiomyopathy, 18 per- 1,000; and in Ibadan, Nigeria, the incidence was 3 per 1,000 cent had hypertensive heart disease, 13 percent had pericar- among children (Hutt 1990). dial disease, and 2 percent had ischemic heart disease (Oyoo A Ghanaian study reported the most common rheumatic and Ogola 1999). valvular lesion to be mitral regurgitation (Amoah 2000). In RHD remains a major cause of heart failure in Africa, a survey of 1,115 children in Kenya, 3 had clinical and especially in the young, and hypertensive heart failure is echocardiographic evidence of RHD, giving a prevalence common, unlike in developed countries, where improved rate of 2.7 per 1,000 (Anabwani and Bonhoeffer 1996), blood pressure control has reduced the prevalence of this whereas 6.2 percent had trivial mitral regurgitation; 0.3 per- condition (Mendez and Cowie 2001). However, as African cent, trivial aortic regurgitation; and 0.4 percent, isolated countries go through the epidemiological transition and mild to moderate regurgitation of the pulmonary valve. develop socioeconomically, the epidemiology of heart fail- Congenital heart disease was found in two children, one ure becomes increasingly similar to that of Western Europe with secundum atrial septal defect and the other with a and North America, with CHD being the most common Cardiovascular Disease | 311 cause of heart failure. Preventive and public health strate- disease is found in a broad swath across Africa, between the gies need to take cognizance of the local epidemiological Sahara Desert in the north and the Zambezi River in the characteristics. south. In temperate climates the disease is rare, presenting Risk factors for cardiac failure in Brazzaville, Republic of as Loeffler's endocarditis, a syndrome exhibiting marked Congo, from 1975 to 1999 were found to be arterial hyper- eosinophilia. Consequently it has been suggested that EMF tension (53 percent), hypercholesterolemia (38 percent), may be caused by an eosinophilic response to filariasis smoking (28 percent), obesity (24 percent), and diabetes or malaria. Another hypothesis from Uganda implicates (5 percent) (Kimbally-Kaky and Bouramoue 2000). Risk cassava protein in the etiology of EMF, but the geographi- factors for ischemic cardiopathy at the Hôpital Principal in cal spread of the disease compared with areas where cassava Dakar, Senegal, were hypercholesterolemia (56 percent), is eaten does not support this as a sole etiological agent tobacco smoking (44 percent), arterial hypertension (Hutt 1990). (41 percent), diabetes (40 percent), and overweight (27 per- In a Kenyan study, patients with echocardiographically cent) (Thiam et al. 2000). Seventy-five percent of patients proven EMF recruited between 1993 and 1996 were found presented with coronary pain, and 50 percent had symp- to have eosinophilia of more than 500 cells per microliter, toms of cardiac insufficiency. compared with non-EMF cardiac patients, and general medical outpatients (Mayanja-Kizza et al. 2000). High eosinophilia of more than 1,000 cells per microliter was Dilated Cardiomyopathy found in 38 percent of the EMF patients but only 6 percent Dilated cardiomyopathy (DCM) resulting in congestive of the non-EMF cardiac patients and in 5 percent of general cardiac failure is surprisingly common in Sub-Saharan medical patients. High levels of eosinophilia in the range of Africa, accounting for up to 20 percent of cardiac cases in the hypereosinophilic syndrome (HES; over 1,500 cells per some regions. Occasionally the disease is familial with spe- microliter) were found in 20 percent of EMF patients but in cific candidate genes recently identified (Sliwa, Damasceno, only 2 percent of the non-EMF cardiac patients and in 1 and Mayosi 2005). A similar picture can be seen in beriberi percent of the general medical patients. The prevalence of and alcoholic cardiac disease. There may be a whole spec- blood and stool parasites was identical in all groups and trum of causes of DCM, including genetic etiology, toxins, could not explain eosinophilia. More often than not the and vitamin or micronutrient deficiency, such as selenium eosinophilic cells in EMF patients were abnormal. deficiency. Dilated cardiomyopathy is often seen as a late complication of human immunodeficiency virus (HIV) HIV-Related Cardiomyopathy infection (Fauci and Lane 2001). In a Ghanaian study of 708 patients in a cardiac referral People with AIDS have evidence of cardiac involvement at center, DCM was the most common form of cardiomyopa- postmortem (40 percent) and by echocardiography (25 per- thy, followed by hypertrophic cardiomyopathy and endomy- cent) (Fauci and Lane 2001). However, fewer than 10 per- ocardial fibrosis (Amoah 2000). Treatment involves nonspe- cent ever experience symptoms. Cardiac involvement is the cific management of the congestive cardiac failure, although cause of death in only 1 to 2 percent of patients infected cardioselective beta-blockade has been increasingly used. with HIV (Boon 2003). These figures may be higher in End-stage disease often necessitates cardiac transplantation, African populations, but good quality epidemiological data although this is rarely feasible in most African countries are lacking. The possibility of higher figures rests on the because of resource constraints. premises that nonischemic cardiomyopathy is more com- mon in African populations than in Western countries and that the etiological factors that cause this high incidence of Endomyocardial Fibrosis cardiomyopathy, be they nutritional deficiencies, genetic Endomyocardial fibrosis (EMF) was first described in detail predispositions, or toxic factors, act synergistically with the by J. N. P. Davies in Uganda in 1948 (Davies 1948). Fibrosis effects of HIV infection on the heart. of the inflow tracts of the right and left ventricles results in HIV infection causes not only cardiomyopathy and heart mitral or tricuspid incompetence and impaired ventricular failure but also CVA, pulmonary embolism, tuberculous function, as a result of the restrictive deficit in which the stiff pericarditis, marantic (nonbacterial) endocarditis, auto- fibrotic ventricles cannot contract and relax normally. nomic dysfunction, and proarrhythmic drug effects. In a Patients usually present in their twenties or thirties. The recent review of 17 peer-reviewed publications covering 312 | Anthony Mbewu and Jean-Claude Mbanya January 1980 to February 2003 on cardiac involvement in mal heart, 66 percent had dilated cardiac cavities and a HIV-infected people living in Africa, Magula and Mayosi hypocontractile left ventricle, and 9 percent had nonob- (2003) showed that cardiac abnormalities are more com- structive cardiomyopathy, dilated cardiac cavities, and a mon in HIV-infected people than in normal controls and hypocontractile left ventricle (Kokou et al. 1999). A study of that about half of hospitalized patients and a significant the left ventricular systolic function of patients with sickle- proportion of patients followed over several years develop cell anemia at the University College Hospital, Ibadan, cardiac abnormalities. The most common HIV-related car- Nigeria, revealed that although the left ventricular mass diac abnormalities were cardiomyopathy and pericardial index was significantly larger in the patients than in the con- disease. Tuberculosis was the major cause of large pericar- trols, there were no significant differences in the left ventric- dial effusion in Africa. HIV-related pericardial effusions are ular systolic function at rest between patients with sickle- usually exudates and tend to occur in patients with cell anemia and age- and sex-matched normal controls advanced disease, with an annual incidence of 10 percent (Adebiyi, Falase, and Akenova 1999). The prominent car- among people living with HIV and AIDS. They are an inde- diovascular abnormalities seen in patients with sickle-cell pendent risk factor for early death, with median survival of anemia, therefore, are most likely to have resulted from less than six months after occurrence of the effusion. left ventricular diastolic dysfunction. Further studies are Myocarditis was the most common pathological abnor- required to evaluate the left ventricular diastolic function in mality in HIV-associated cardiomyopathy, and nonviral patients with sickle-cell anemia as well as their cardiac func- opportunistic infections, such as toxoplasmosis and tion during exercise and during episodes of crisis. Cryptococcus, may account for up to 50 percent of such cases in Africa. Although the mechanisms involved in cardiomy- Congenital Heart Disease opathy in people with HIV infection are poorly defined, a role for direct retroviral action or focal infiltration of acti- Ventricular and atrial septal defect, Fallot's tetralogy, and vated immune cells, or both, has been postulated. Recent patent ductus arteriosus were the most common congenital studies have demonstrated cardiac myocyte protein infiltra- lesions in a Ghanaian study (Amoah 2000). The major tion in AIDS-related cardiomyopathies, rather than focal cardiovascular disorders in children were congenital heart immune cell lesions (Magula and Mayosi 2003). In Africa, disease and RHD. Idiopathic cardiomyopathy was rare. In however, other factors, such as nutritional deficiencies and a study of 13,322 schoolchildren from Sahafa Town, the cardiotoxicity of antiretroviral medications, are asso- Khartoum, Sudan, from 1986 to 1990, the prevalence of con- ciated with dilated cardiomyopathy in HIV-infected patients genital heart disease was 2.0 per 1,000, with ventricular sep- (Magula and Mayosi 2003). Long-term cardiac side effects of tal defect, atrial septal defect, patent ductus arteriosus, and antiretroviral therapy are likely to become increasingly com- Fallot's tetralogy making up 85 percent of the cases (Khalil mon in Africa as antiretroviral therapy becomes more read- et al. 1997). Patent ductus arteriosus and atrial septal defect ily available through national treatment programs, such as were twice as common in females as in males. The preva- the Plan for the Comprehensive Treatment and Care of HIV lence rate was comparable to that of similar African coun- and AIDS in South Africa (http://www.doh.gov.za), the tries but lower than European and North American rates. WHO "3 by 5" Initiative (WHO 2004), and the President's Emergency Plan for AIDS Relief sponsored by the United Dysrhythmias States (IOM 2005). Most African programs, however, do not include protease inhibitors, responsible for much of Atrial fibrillation is the most common cardiac arrhythmia in the long-term cardiac toxicity, in their first regimen. South Africa, often related to RHD, and is responsible for Antiretroviral cardiac side effects may become apparent only significant morbidity and mortality in the general popula- in several years time as protease inhibitors come off patent tion. The incidence of atrial fibrillation in South Africa is and become affordable to national programs in Africa. about 8 percent of the population 70 years and older. Atrial fibrillation affects more than 5 percent of the population over 65 and 10 percent of those over 80. The incidence is Sickle-Cell Disease higher in patients with left ventricular hypertrophy and A study of 70 children between the ages of 3 and 16 years heart failure. Intraventricular malconductions seem to with homozygous sickle-cell anemia in Lomé, Togo, from exhibit a prevalence in African populations similar to that in January 1996 to April 1997 found that 26 percent had a nor- other parts of the world (Omotoso and Kane 2000). Cardiovascular Disease | 313 Pericarditis Genetic determinants provide the foundation on which behavioral, sociocultural, economic, and educational deter- A study carried out from 1989 to 1996 in the Republic of minants build. As an example, the popular hypothesis that Congo found that 4.9 percent of patients with cardiovascu- African people have a genetic predisposition for salt reten- lar disease had nonrheumatic pericarditis with effusion tion may be compounded in urban African cultures in (Nkoua, Tsombou, and Bouramoue 1999). Twenty-two which salt intake is high, thus causing hypertension and in percent were HIV positive. The principal cause of the peri- part explaining the documented higher incidence of hyper- carditis was tuberculosis, which accounted for about a quar- tension in urban than rural African societies (Steyn and ter of the patients, all of whom were HIV positive; all those Fourie 1991). with benign acute pericarditis were HIV negative; and all Risk factors arise from determinants and are directly those with lymphocytic pericarditis were HIV positive. Of linked to a disease in a causal fashion, although not every- those affected by cardiac tamponade, just under a third were one with the risk factor develops the disease. Risk factors HIV positive. The mortality rate was 11.0 percent--9.9 per- include smoking, high blood pressure, malnutrition, obesity, cent among those who were HIV negative and 15.0 percent hyperlipidemia, lack of physical exercise, and beta-hemolytic among those who were HIV positive. This study confirms streptococcal infection. Many of the determinants of CVD the high rate of nonrheumatic pericardial effusion, the role are shared with cancer, diabetes, and chronic obstructive of HIV infection, and the leading place of tuberculosis disease. Table 21.3 illustrates the relation between CVD among causes. These findings corroborate those suggesting determinants, risk factors, behaviors, and disease (Reddy that the outcome of pericardial effusion associated with 2004). HIV infection can be cardiac tamponade (Nkoua, Tsombou, Coronary heart disease and stroke share the same risk and Bouramoue 1999). factors, but their relative implication in the occurrence of these diseases is different. Since the middle of the twentieth DETERMINANTS, BEHAVIORS, century the prevalence of all risks factors for CVDs, except AND RISK FACTORS hyperlipidemia, has been increasing. Hyperlipidemia, although less prevalent than in developed countries, is In the past few decades appreciation of the importance of found in patients with metabolic disorders and families with determinants and risk factors in the etiology of CVD has genetic susceptibility (Law 1998). Risk factors for CVD pres- grown. Determinants are the ecological factors that provide ent in developed countries are the same in Sub-Saharan the milieu in which a disease develops and need not be Africa, but their association and the possibly differing genetic directly linked to the disease causally. Most people exposed susceptibilities may be responsible for the particular pattern to the determinant do not inevitably develop the disease. of CVD in Africa. Table 21.3 The Relation between CVD, Risk Factors, Behaviors, and Determinants Disease or disorder Risk factors Behaviors Determinants Cardiovascular disease Tobacco Smoking tobacco Psychosocial (stroke, ischemic heart disease, Chewing tobacco Educational peripheral vascular disease) Alcohol Alcohol misuse Environmental Food Food consumption Economic eating Commercial cooking purchasing Advertising Hypertension Salt use Marketing Health care­seeking behavior Food labeling Medication compliance behavior Obesity Sedentary behavior Lack of physical activity Source: Reddy 2004. 314 | Anthony Mbewu and Jean-Claude Mbanya Table 21.4 Risk Factors for CHD Reported from Hospital Patients Age (mean Sample Hypertension Diabetes Obesity Smoking Author and date or range) size (%) (%) (%) (%) Bertrand 1992 45 100 25.8 12.9 29 60.0 Steyn and Fourie 1991 15­64 986 14.4 (M) -- -- 22.0 (M) 13.7 (F) 8.4 (F) Rosman 1986 20 -- 69.8 -- -- -- Source: Compiled by authors. Note: -- not available; M male; F female. A Tunisian population study estimated the prevalence of individuals. Nevertheless, communities still exist in the risk factors for CVD to be 19 percent for hypertension, Democratic Republic of Congo, Kenya, Nigeria, and the 10 percent for diabetes, 28 percent for obesity, 36 percent for Kalahari Desert in which blood pressure is low and does not android obesity, and 21 percent for smoking (Ghanem and seem to rise with age. Rural-to-urban migration coupled Fredj 1999). Total calorie intake was 2,483 kilocalories, com- with acculturation and modernization trends have some prised of 67 percent carbohydrates, 18 percent protein, relation to the development of high blood pressure as and 15 percent fat. Among urban dwellers in Sub-Saharan observed in Kenyan and Ghanaian epidemiologic studies Africa, intakes of food, especially fat, have risen, and intakes (table 21.5). of high-fiber foods have fallen. The mean serum cholesterol The prevalence of hypertension is particularly high in level is significantly higher than that of rural populations urban settings in Sub-Saharan Africa; between 8 and 25 per- living traditionally (Steyn and Fourie 1991). Obesity in cent of the adult population are affected, depending on what females has risen enormously. The prevalence of hyperten- definition of hypertension is used. The two commonly used sion exceeds that in developed countries. The same applies are the Joint National Committee on Detection, Evaluation, to the practice of smoking in males but not in females. The and Treatment of High Blood Pressure VI (JNC VI) defi- level of physical activity has fallen. Table 21.4 shows the nition (JNC 1997), which is a systolic pressure above 140 relative prevalence of the most important risk factors for and diastolic pressure above 90 millimeters of mercury coronary heart disease in hospital patients. (mmHg), and a more conservative cutoff of 160 systolic Stroke and coronary heart disease occurs earlier in the pressure and 100 diastolic, as used in many African control lives of people in Africa, as in other low-income countries, programs. Over 80 percent of hypertensive patients in clini- than in the industrial world. Whereas CVDs are diseases of cal practice have essential hypertension (that is, primary elderly people in developed countries, where they occur hypertension with no known cause), with most of the after the age of 60, they are preponderant in Africans even remainder having a renal origin for their hypertension before the age of 40. (Akinkugbe 1976). Before the latter half of the twentieth century most people in most Sub-Saharan Africa countries did not live Hypertension beyond 40 years, the age at which hypertension becomes At the beginning of the twentieth century, high blood pres- increasingly more prevalent. Some earlier researchers sure was virtually nonexistent among indigenous Kenyans suggested that Africans did not show the characteristic (Lore 1993) and Ugandans (Hutt 1990), but the reason may increase in blood pressure with age (Shaper 1974), but this have been the lack of screening programs and access to care. observation may have been due to deficiencies in the design From about 1975, high blood pressure became established of cohort studies. in Cameroon, Côte d'Ivoire, Democratic Republic of Congo, So prevalent is hypertension today in Sub-Saharan Africa Ghana, Kenya, Nigeria, and Uganda. As in developed coun- that hypertensive heart disease might in fact be the most tries, consumption of salt and alcohol, psychological stress, common form of CVD in Africa. Hypertension is a risk obesity, physical inactivity, and other dietary factors are factor for both stroke and IHD (Bradshaw et al. 2003). Left thought to have played an important etiologic role in the ventricular hypertrophy, congestive heart failure, and stroke genesis of primary hypertension in genetically predisposed are common in Africans with hypertension. There is little Cardiovascular Disease | 315 Table 21.5 Prevalence of Hypertension, by Country Prevalence (%) Country Author and date Criteria Population Male Female Cameroon Mbanya et al. 1998 (1) Urban 16.4 12.1 Cameroon Mbanya et al. 1998 (1) Rural 5.4 5.9 Mauritius Nan et al. 1991 (1) Urban and rural 11.0 2.0 Nigeria Cooper, Rotimi, Ataman et al. 1997 (2) Urban and rural 15.0 14.0 Senegal Astagneau et al. 1992 (1) Urban 11.0 11.0 Senegal Astagneau et al. 1992 (2) Urban 24.0 22.0 South Africa Steyn et al. 1996 (1) Urban 13.0 2.0 South Africa Metcalf et al. 1996 (1) Rural 16.0 2.0 Tanzania Berrios et al. 1997 (1) Urban 9.0 12.0 Tanzania Edwards et al. 2000 (2) Urban 30.0 28.6 Tanzania Edwards et al. 2000 (2) Rural 32.2 31.5 Source: Compiled by authors. Note: (1) 160/95 mmHg or treated or self-reported; (2) 140/90 mmHg or treated or self-reported. published information on formal programs addressing hypertensive patients confirm that patients are unable to awareness, treatment, and control. Local, regional, and perceive changes in their blood pressure and should be national surveys are required to provide epidemiological taught to rely on regular blood pressure checks by their data necessary for informed decision making and policy physician (Familoni and Ariba 2003). setting on when and whom to treat in Africa (Kapuku, In Western societies, such as the United States and the Mensah, and Cooper 1998; van der Sande et al. 2001). United Kingdom, the prevalence of hypertension and stan- There seem to be marked urban-rural differences in the dardized mortality rates from stroke are higher for people prevalence of the disease (table 21.5). Prevalence levels are of African origin than for whites (Cooper, Rotimi, higher in South Africa among urban Zulu people than Kaufman, et al. 1997). The same pattern is emerging in among their rural counterparts (Mokhobo 1976; Seedat, Sub-Saharan Africa. Thus in South Africa, age-adjusted Seedat, and Hackland 1982) and among urban Xhosa peo- hypertension prevalence and age-specific rates of death ple in Cape Town than their rural relatives in the Eastern from stroke are higher among urban blacks than equivalent Cape (Sever et al. 1980; Steyn et al. 1993). In Cameroon, white populations (Opie and Steyn 1995). There is evidence age-adjusted rates of blood pressure in urban areas were that hypertension is an important cause of mortality in greater than or equal to 160 mmHg systolic or 95 mmHg Sub-Saharan Africa. The Adult Morbidity and Mortality diastolic, and treatment of hypertension rose from 5 per- Project (AMMP 1997) showed that the probability of dying cent in rural areas to 17 percent in urban ones (Cruickshank from an NCD (which is largely made up of stroke in et al. 2001). Tanzania) between the ages of 15 and 60 years is more than Few studies from Africa have reported on hypertension six times higher in an urban area and between two and four treatment and control. In the Black Risk Factors Study times higher in two rural areas in Tanzania than in the (BRISK) study in urban black townships in the Cape penin- United Kingdom (AMMP 1997). In the same study, NCD sula of Cape Town, South Africa, 61 percent of those with was the most common cause of death in the urban area hypertension (greater than or equal to 160 over 95 mmHg) and one of the rural areas for those over 60 years of age were aware of their hypertension, and 48 percent were treat- (figure 21.2). ed. In an urban population in the Democratic Republic of As in other parts of the world, the prevalence of hyper- Congo only 31 percent of those with hypertension (blood tension in the Sub-Saharan Africa region has increased as a pressure greater than or equal to 160 over 95 mmHg) were manifestation of the epidemiological transition (Omran aware of their diagnosis; 13 percent were treated and 3 per- 1983). This implies that, as elsewhere in the world, environ- cent of those with hypertension had their disease controlled mental factors related to urbanization and increasing (M'Buyamba-Kabangu et al. 1986). Nigerian studies of affluence are important determinants of the disease. 316 | Anthony Mbewu and Jean-Claude Mbanya Figure 21.2 Probability of Death by Broad Cause in Men eridemia) or a family history of hypercholesterolemia. It is, between the Ages of 15 and 60 Years in Tanzania however, present in 10 to 70 percent of patients with 20 antecedent CHD or stroke. A study of black African patients 18.2 admitted to a coronary care unit in Cape Town, South Africa, with acute ischemic syndromes and myocardial 15 14.2 infarction revealed relative hyperlipidemia among the 11.1 patients but not in the healthy controls (Mayosi et al. 1997). 10 percent 7.9 5 Diabetes 2 0.2 Diabetes mellitus is a well-established risk factor for CVD 0 infectious diseases noncommunicable diseases injury (King, Aubert, and Herman 1998). The prevalence of type 2 Hai, Tanzania United Kingdom diabetes in Africa is about 2.5 percent, ranging from 0.8 per- cent in rural Cameroon (Mbanya et al. 1997) to 13.5 percent Source: AMPP 1997. in Mauritius (Dowse et al. 1990). Type 2 diabetes is more frequent in South Africa and North Africa than in central As the number of fatalities from cardiovascular diseases and West Africa, and it increases from rural to urban areas declines in Western industrial nations, an opposite trend is (Cooper, Rotimi, Kaufman, et al. 1997; King and Zimmet observed in East Africa (Mbaya 1998). Interregional varia- 1988). The World Bank ranks these countries as upper- tions in the prevalence of vascular disorders have been middle-income countries, and they are further along the attributed to socioeconomic, psychosocial, and heritable epidemiological transition than the low-income countries physiological parameters. of Sub-Saharan Africa (table 21.6). Among the Luo of Kenya, increasing blood pressure within months of migrating from the rural areas to the city Tobacco Smoking and Alcohol Consumption has been recorded, with concomitant increases in their dietary sodium and declines in their dietary potassium Tobacco consumption in South Africa has declined since (Poulter 1988; Poulter et al. 1984). A prospective cross- 1994 due to stringent antitobacco legislation and hikes in sectional study in rural Nigeria showed the prevalence of excise duty, resulting in 25 percent reductions in both obesity to be 2 percent, with 1.2 percent in males and 3.2 the number of people smoking and overall tobacco con- percent in females (Okesina et al. 1999). High blood pres- sumption (Reddy 1997; Steyn 1998) (see table 21.7). Alcohol sure, observed in 15.2 percent of the subjects, occurred more consumption, however, is relatively frequent in Africa. The among males (19.1 percent) than females (10.3 percent). types of alcohol consumed include wine, beer, and locally Malignant hypertension is common. made beverages. Alcohol has been implicated in the devel- Hypertensive patients whose blood pressure fails to fall at opment of hypertension, CHD, stroke, and heart failure night in the normal 24-hour rhythm (nondippers) have a (table 21.8). higher incidence of cardiovascular complication, early glomerular failure, and microalbuminuria. A Nigerian study identified 28 percent nondippers, 57.1 percent of whom had Lack of Physical Activity microalbuminuria (Alebiosu et al. 2004). Physical activity is more prevalent in rural than urban As regards hypertension control, the guidelines of JNC VI regions of Africa, and that partly explains the higher preva- may be more relevant to Sub-Saharan Africa than JNC VII lence of obesity in urban areas. The prevalence of sedentary because of the difficulty of effecting even moderate control lifestyles in Cape Town, South Africa, among individuals of hypertension in Sub-Saharan Africa. 30 years of age and above was 39 percent for men and 44 per- cent for women (Levitt et al. 1993). Twenty-two percent of men and 52 percent of women in urban Tanzania (Edwards Hyperlipidemia et al. 2000) had low levels of physical activity. Also, 10 per- Hyperlipidemia is uncommon in Africa, being present cent of men and 15 percent of women in rural areas mainly in patients with metabolic disorders (hypertriglyc- indicated low physical activity during the same study. Cardiovascular Disease | 317 Table 21.6 Prevalence of Diabetes, by Country Country Author and date Method Site Sample size Prevalence (%) Cameroon Mbanya et al. 1997 Blood Urban 1,048 2.8 Cameroon Mbanya et al. 1997 Blood Rural 719 1.1 Ethiopia Peters 1983 Urine Urban and rural 2,381 0.3 Lesotho Politzer and Sachs 1967 Urine Rural 3,000 0.2 Malawi Davidson 1963 Urine Urban and rural 4,725 0.1 Mali Fisch et al. 1987 Blood Rural 7,472 0.9 Nigeria Ohwovoriole, Kuti, and Kabiawu 1988 Blood Urban 1,627 1.7 Nigeria Erasmus et al. 1989 Blood Urban and rural 2,800 1.4 Nigeria Owoaje et al. 1997 Blood Urban 247 2.8 Sierra Leone Ceesay et al. 1997 Blood Urban 501 2.4 Sierra Leone Ceesay et al. 1997 Blood Rural -- 0.0 South Africa Levitt et al. 1993 Blood Urban 729 8.0 South Africa Omar et al. 1993 Blood Urban 479 5.3 South Africa Mollentze et al. 1995 Blood Urban 758 6.0 South Africa Mollentze et al. 1995 Blood Rural 853 4.8 Tanzania Swai et al. 1993 Blood Urban 1,255 8.8 Togo Teuscher et al. 1987 Blood Rural 1,381 0.0 Source: Compiled by authors. Note: -- not available. Table 21.7 Prevalence of Tobacco Smoking, by Country Prevalence (%) Country Author and date Population Men Women Ethiopia Betre, Kebede, and Kassaye 1997 Urban 11.8 1.1 Tanzania Bovet et al. 2002 Urban 22.0 2.6 Tanzania (Kilimanjaro) Swai et al. 1993 Rural 42.6 2.1 Tanzania (Morogoro) Swai et al. 1993 Rural 28.2 3.9 Tanzania (Mara) Swai et al. 1993 Rural 8.6 2.7 Source: Compiled by authors. Table 21.8 Prevalence of Alcohol Consumption, by Country Prevalence (%) Country Author and date Population Men Women Tanzaniaa Edwards et al. 2000 Urban 6.1 2.5 Tanzaniaa Edwards et al. 2000 Rural 3.0 0.0 Ethiopia Betre, Kebede, and Kassaye 1997 Urban 34.0 34.0 Nigeria Kadiri and Salako 1997 Urban 5.4 2.8 Zimbabweb Chinyadza et al. 1993 Urban 63.0 41.0 Source: Compiled by authors. a. Heavy drinkers (more than 49 units per week for men and more than 35 units per week for women). b. Hospital-based study. 318 | Anthony Mbewu and Jean-Claude Mbanya Table 21.9 Prevalence of Obesity, by Country Prevalence (%) Country Author and date Population Male Female Cameroon Sobngwi et al. 2002 Urban 17.1 3.0 Cameroon Sobngwi et al. 2002 Rural 1.2 5.4 Gambia, The van der Sande et al. 2001 Urban 4.0 4.0 Malawi Msamati and Igbigbi 2000 Urban 0.0 11.4 Tanzania (Mara) Swai et al. 1993 Rural 0.9 0.8 Tanzania (Kilimanjaro) Swai et al. 1993 Rural 0.1 1.5 Tanzania (Morogoro) Swai et al. 1993 Rural 0.1 1.1 Tanzania (Ilala) Edwards et al. 2000 Urban 6.8 17.2 Tanzania (Shari) Edwards et al. 2000 Rural 3.2 4.8 Tanzania (Dar es Salaam) Bovet et al. 2002 Urban 6.9 17.4 Source: Compiled by authors. Note: Obesity defined as body mass index greater than 30 kg/m2. Obesity events, but the association is likely to be weak, as shown in one study of myocardial infarction in young South African Obesity is increasing in prevalence in Sub-Saharan Africa; Indians (Ranjith et al. 2004). particularly among urban women (tables 21.9 and 21.10). However, the value of the waist-hip ratio has never been properly validated in large population-based studies in Sub- PREVENTION, REHABILITATION, Saharan Africa in regard to their predictive value for cardio- AND CURE OF CARDIOVASCULAR DISEASE vascular outcomes. The same holds true for waist cir- cumference of greater than 88 centimeters for females and Traditionally, interventions have focused on altering risk 102 centimeters for males. factors by, for example, eradicating streptococcal infection using antibiotics in order to prevent rheumatic fever and subsequent RHD. The health promotion approach, however, Genetic Determinants teaches that to be fully effective in curing, ameliorating, or The renin-angiotensin system and associated gene polymor- preventing a disease, one must understand how the deter- phisms may be important in predicting cardiovascular minants give rise to a disease within particular settings. One can then develop determinant-based interventions. Thus, Table 21.10 Prevalence of Obesity in Women Age 15 to for example, rheumatic fever typically develops in situations 49 Years, by Country of poverty and overcrowding, where children are malnour- ished, immunocompromised, and susceptible to bacterial Prevalence (%) infection. An intervention that focuses on providing peni- Country Sample size Urban Rural cillin for those children with streptococcal sore throat will Benin 2,266 3.5 1.4 not be fully effective in eradicating the disease. The determi- Burkina Faso 3,161 3.5 0.6 nants of the disease must also be dealt with by educating the Central African Republic 2,025 2.0 0.5 caregiver to appreciate the need for prompt treatment of Côte d'Ivoire 3,108 6.2 1.3 sore throats, educating health professionals to take a throat Ghana 1,773 8.1 1.5 swab before starting antibiotic treatment, providing child Namibia 2,205 13.4 3.4 support grants to caregivers so that they can buy food for Niger 3,292 6.4 0.3 the malnourished child, alleviating overcrowding so that the Senegal 2,895 7.2 1.9 sick child does not sleep six to a bed with his or her siblings, Zimbabwe 1,968 12.5 3.4 thus spreading infection, and so on. Health promotion also teaches the importance of the Source: Martorell et al. 2000. Note: Obesity defined as body mass index greater than 30 kg/m2. "settings approach" to disease prevention and control. It Cardiovascular Disease | 319 appreciates that the bulk of disease management does not IMPACT ON THE HEALTH CARE SYSTEM IN take place in the hospital or clinic but in the home, the AFRICA AND STRATEGIES FOR CONTROL school, and the workplace. Understanding the etiology, AND PREVENTION natural history, and management paradigms of the disease within these settings is therefore crucial to developing Options to improve monitoring of CVD morbidity and prevention, rehabilitation, and cure programs. risk-factor levels include establishing community registries These interventions may be tailored for the population, of stroke and repeatedly examining representative popula- such as screening schoolchildren by taking throat swabs for tion samples. Nationwide vital registration to monitor CVD beta-hemolytic streptococcus after an outbreak of rheu- mortality would be desirable but appears not to be feasible matic fever in a school or treating an infected child with at present, and maintaining a sample registration system penicillin for a streptococcal sore throat. Primary preven- would be prohibitively expensive. tion may include the prevention and treatment of strepto- In spite of the current low prevalence of hypertension in coccal infection by providing school feeding programs for some countries, the total number of people with hyperten- malnourished children and providing penicillin to treat sion in the developing world is high, and a cost analysis of sore throats or training health personnel in the diagnosis of possible antihypertensive drug treatment indicates that the condition and educating the caregiver. Secondary pre- developing countries cannot afford the same treatment as vention is treating with penicillin for several years after an developed countries (Seedat 2000). Only 20 percent of the attack of rheumatic fever in order to prevent rheumatic hypertension in the United States is under control, whereas heart disease; and tertiary prevention could involve treating in Africa only 5 to 10 percent is controlled at a blood pres- the valvular cardiac disease of RHD symptomatically with sure of less than 140 over 90 mmHg (Pretoria Department diuretics and digoxin, curatively with balloon valvuloplasty of Health 2002). or cardiac surgery. It is claimed that black patients respond well to thiazide A modern understanding of disease management empha- diuretics, calcium channel blockers, and vasodilators such sizes the importance of analyses of the cost-effectiveness of as alpha-blockers, hydralazine, and reserpine, and respond disease interventions in order to build sustainable systems poorly to beta-blockers, angiotensin-converting enzyme for disease control. This is important for developed coun- inhibitors, and angiotensin II receptor antagonists unless tries as well as for developing countries. It is particularly they are combined with a diuretic. This is increasingly being critical for the prevention and control of cardiovascular brought into doubt, however. Comprehensive cardiovascu- disease, for which secondary and tertiary interventions are lar disease research programs in Africa are needed, as social, often extremely expensive and require high-grade technical economic, and cultural factors impair control of hyperten- skills to administer. A difficult tradeoff has to be made in sion in developing countries. Hypertension control is ideally such CVD control programs between the different interven- suited to the initial component of an integrated CVD con- tions, and this is best done rationally rather than emotion- trol program. Primary prevention, through a program ally. The tradeoff may be between providing antibiotics to focused on lifestyle should be synergistically linked to cost- prevent rheumatic fever versus providing surgery to babies effective methods of detection and management. The exist- with congenital heart disease, a difficult choice to make. The ing health care infrastructure needs to be oriented to meet former intervention, however, may be far more cost-effective the emerging challenge of CVD while empowering the com- in the long run. munity through health education. It is extremely difficult for policy makers and health plan- The lipid cardiovascular risk-factor profile of African ners to articulate and make these choices, as they often come populations continues to be relatively benign, with mean with considerable political as well as emotional cost. Often total cholesterol, LDL cholesterol and HDL cholesterol, and the hard choice is masked under a "waiting list" on which median triglyceride levels in the "low risk" ranges, even children may die without an overt decision having been among cohorts with established CVD. The other major risk made to deny them lifesaving interventions. In Sub-Saharan factors, however, are significantly elevated in urban popula- Africa there often is no choice to make because the suffering tions, particularly prevalent hypertension, increasing and dying patients never even reach the clinic to demand obesity, and type 2 diabetes mellitus, as well as growing treatment, or the facilities and surgeons to provide the rates of tobacco use. Furthermore, the prevalence of risk surgery are nonexistent. factors has risen in both urban and rural areas. A 1998 study 320 | Anthony Mbewu and Jean-Claude Mbanya carried out by the WHO in Tanzania found that hyper- focus on high-risk groups, and prevention needs a multisec- cholesterolemia (total cholesterol greater than 5.2 millime- toral, health promotion approach. ters molar per liter) was present in 21.8 percent of men A suitable strategy to adopt in the prevention and control and 54 percent of women; that measured weight, body mass of NCDs is the WHO Global Strategy for Noncommunicable index (BMI), and prevalence of obesity (BMI greater than Diseases, whose principal objectives are to do the following: or equal to 30 kilograms per square meter) had increased significantly to 22.8 percent among women in urban · Map the emerging epidemics of NCD and analyze their Dar es Salaam; and that the overall prevalence of hyper- determinants. tension (blood pressure higher than 160 over 95 mmHg · Reduce the levels of exposure of populations to common or antihypertensive drug use) was 41.1 percent (Njelekela risk factors and their determinants. et al. 2001). · Develop norms and guidelines for effective interventions. These classical risk factors operate in much the same way as in northern European populations with regard to NCDs, including cardiovascular and cerebrovascular being predictive of CVD (Kruger, Venter, and Vorster 2001; disease, pulmonary diseases, liver disease, cancer, diabetes, Yusuf et al. 2004). To date no evidence has been found that osteoporosis, and trauma, constitute the major cause of Africans have esoteric, genetically determined risk factors death in developed countries and are predictably emerging that differ from other populations in the world; nor do they as significant threats to health in countries at intermediate appear to have special genetic traits that protect them from stages of the epidemiological transition (Yusuf et al. 2004). heart disease. It seems likely that the continuing socioeco- Interventions could be designed based on the philosophy nomic advance in many African countries will be accompa- that diseases with common risk factors, such as smoking, nied by the attendant increases in CVD that are seen in increased fat and salt in the diet, alcohol abuse, and so on, epidemiological transitions across the world. The challenge require common preventive strategies. is to introduce primary and secondary prevention measures Screening programs to detect hypertension in the com- now, before the epidemic of CVD accelerates, particularly as munity may be appropriate, as the disease does not become such strategies may be more cost-effective than angioplasty clinically manifest until it causes end organ damage, such as and cardiac surgery in the cash-strapped economies of heart failure, heart attack, stroke, and kidney failure. Such Sub-Saharan Africa. measures would be subject to the constraints of available Up to 22 percent of premature all-cause mortality and resources, however; and an approach that targets high-risk 45 percent of stroke mortality could be reduced by appro- patients, that is, those who present with other related condi- priate detection and treatment (Yusuf et al. 2004). Cheap, tions or with multiple risk factors, may be more feasible effective therapy is available. With mortality risk now higher than populationwide screening in poorer African countries from NCDs than from communicable diseases in Sub- (WHO 2002). Saharan Africa and elsewhere, systematic measurement, Health systems in developing countries will need to be detection, and genuine control of hypertension once treated drastically improved to deal with the burden, as they often can go hand-in-hand with other adult health programs in suffer from inefficient allocation of resources and manage- primary care. In many African countries undernutrition ment of services. Better management must be based on coexists with obesity in the same communities, demonstrat- both the high-risk-behaviors approach, such as detection ing a double burden of disease, such as in The Gambia, and management of obesity, hypertension, and diabetes at where a 1996 community-based survey demonstrated a the primary health care level, and the populationwide prevalence of undernutrition (BMI less than 18 kilograms approach, such as education through the mass media and per square meter) of 8 percent, whereas prevalence of schools. Integration of the package of CVD prevention and obesity (BMI greater than or equal to 30 kilograms per control within the programs of ministries of health will be square meter) was 4 percent. However, obesity was higher needed. (32.6 percent) among urban women 35 years or older (van Increasing availability of drugs and diagnostics is a major der Sande et al. 2001). Furthermore, CVD risk factors tend issue in Africa, as it was for HIV/AIDS and tuberculosis. to cluster, with the potential for synergistic effects, as has Local manufacture of off-patent CVD drugs may improve been described in other populations in the world (Kruger, access as well as stimulate the economy, as proposed in Venter, and Vorster 2001). Differential interventions should the Initiative for Pharmaceutical Technology Transfer of Cardiovascular Disease | 321 South Africa, a program of the New Partnership for Africa's The data are scanty and there is a dire need for more health Development. economics research. Health technology for management of these CVDs will Development of a global program to provide access to need to be appropriate, affordable, and sustainable and not CVD drugs, similar to the "3 by 5" Initiative of the WHO for simply imported from the West. The detection, prevention, antiretroviral drugs, is unlikely, because cardiovascular and control of CVD in women is a major issue in Sub- deaths do not have the same emotive power as AIDS deaths. Saharan Africa, where women may suffer inferior diagnosis Yet CVD deaths are often sudden, striking down victims in of CVD and limited access to care because of adverse their most productive years and resulting in deep emotional gender-power relations between men and women. and economic shock for their families. For survivors, the CVD control in Africa will probably not reach a high level years lived with disability, such as heart failure or hemiplegia, of adequacy for a few decades to come, given the scant are often traumatic and expensive because of health care human, financial, and infrastructural resources in its coun- costs and lost productivity. The prognosis for living with tries. However, those countries cannot ignore the problem if heart failure is like that for someone living with HIV/AIDS they hold fast to the dictum of "health as a human right." on antiretrovirals; the annual mortality is 5 to 15 percent. They will, however, have to carefully categorize and priori- Perhaps the health systems being set up in African countries tize their public health expenditure on the prevention and for provision of antiretroviral drugs and comprehensive control of CVD. The experience of developed countries, AIDS care could have the beneficial side effect of providing who are further on in the course of their "epidemiological systems and personnel for CVD surveillance and care. transition," suggests that primary prevention of CVD now is In many African countries medicinal plants are used to likely to be more cost-effective than acute care for affected treat cardiovascular complaints (Diallo et al. 1999). If these patients in 20 years. In the United States, for example, the are truly effective, it could in part explain the low prevalence direct costs of CVD are estimated to account for 2 percent of of CVD (and cancer) in rural African populations. the gross domestic product. In Canada, CVD accounts for Indigenous knowledge systems should be researched, in 21 percent of total disease-classifiable costs of illness. The order to determine whether to popularize natural medicines data for other chronic diseases in developing countries sug- for the prevention and cure of CVD. Cardiovascular disor- gest that these scales of health care costs are likely to be sim- ders currently receive little or no attention in most African ilar. As an example, in India, the annual cost of drugs for a countries (Muna 1993b). Projections based on recent stud- patient with type 2 diabetes mellitus is US$70, considerably ies suggest that the management of these disorders will rep- more than the annual per capita expenditure on health care in resent a major challenge for the overextended and shrinking that country. In Tanzania, diabetes care swallows up 8.1 per- health budgets of these poor nations in the near future. cent of the health care budget (http://www.ichealth.org). Given the prevalence of such common cardiovascular The costs of intersectoral, primary preventive measures conditions as hypertension, which in some cases can be to control CVD are likely to be much lower than those for 25 percent or higher, treatment is beyond the budgetary acute, secondary care. As an example, the cost-effectiveness possibilities of any of the African countries. Other condi- of introducing antitobacco legislation, or limiting the quan- tions, including infections involving the heart and related tities of salt permitted in processed food, is likely to be sub- structures, rheumatic fever and its complications, cardiomy- stantially higher than building coronary care units. The evi- opathies, and congenital heart disease, are also common. dence from developed countries since 1990 suggests that Recent trends suggest that ischemic heart disease may be less most of the decline in CVD mortality in these countries has uncommon than was previously thought. Health planners resulted from better secondary care for CVD, but the relative and policy makers must be educated on the crucial role of cost per life saved is still high, too high for health systems in health research in general, and cardiovascular disorders developing countries to cope with. Developed countries research in particular, as a basis for formulating a rational faced the peak in their incidence of CVD when they were health care policy and making managerial decisions. The already wealthy and had the resources to cope. Thus, the need for training and funding, and especially the leadership stratagems adopted by developing countries are likely to be required to develop and sustain research activity, will different. In secondary care, African countries of the south require a multidisciplinary, multidirectional, collaborative will need to develop essential vascular packages, such as approach at national and international levels, as well as firm aspirin and beta-blockers for patients with angina pectoris. local commitment. As is the case for most other important 322 | Anthony Mbewu and Jean-Claude Mbanya health problems, cardiovascular disorders are rapidly some way immune from CVD has been dispelled by well- becoming a global issue and should be recognized as such. documented evidence from, among other countries, Following the guidelines of JNC VI in the screening, Cameroon, Caribbean countries, Mauritius, the Seychelles, detection, and treatment of all hypertensive patients is South Africa, Tanzania, and the United States. In such coun- beyond the health budgets of most African countries tries CVD already accounts for between 15 and 40 percent of (Gaziano and Opie 2001). Most countries, therefore, adopt all deaths. an opportunistic approach to the identification of patients Data on the burden of cardiovascular disease for most with hypertension and aim for significantly higher target African countries are, however, scanty. The estimations blood pressures than those recommended. The result is that made in the Global Burden of Disease Study for CVD in considerable numbers of patients continue to suffer hyper- 1990 in Africa relied on data from only 2 percent of the 700 tension-related strokes leading to disability or death. This million inhabitants of Africa (Murray and Lopez 1996). Still, inadequate detection and treatment of hypertension is the studies do suggest that the prevalence of risk factors for most likely explanation for the high proportion of hemor- CVD, and the behavioral, socioeconomic, and demographic rhagic strokes in African populations. determinants that underlie these risk factors, are increasing The major challenges in the management of hyperten- in Africa, despite the persistence of an appalling burden of sion are therefore twofold. First is to develop cost-effective, communicable disease and diseases of poverty. Further- population-based, health promotive strategies for primary more, the prevailing pattern of past decades, when CVD in prevention, and secondary amelioration of hypertension. Africa was predominantly RHD, hypertensive heart disease, The second challenge is to lower the costs of treatments stroke, and dilated cardiomyopathy, is now beginning to through strategies such as the local manufacture of antihy- change, with an increasing proportion of morbidity and pertensive drugs that have come off patent (IPTT 2003). mortality due to ischemic heart disease. These are also probably the principal challenges in the man- It is probably cheaper and more effective to take popula- agement of cardiovascular disease in Africa, as hypertensive tion-based measures to prevent heart disease and stroke stroke is the leading cause of cardiovascular death and than it is to treat the people with surgery and drugs once disability in Africa. Research is under way to develop these CVD is established. RHD and heart failure also have highly prevention and control strategies (http://www.ichealth.org). effective primary and secondary interventions that can be Research continues into which of the two approaches to readily applied. The time to act in prevention and control of hypertensive control is most cost-effective: treating hyper- CVD is now. tension according to the absolute risk for CVD or following an arbitrary blood pressure target. The former is likely to be more cost-effective, but more research is still needed to settle REFERENCES the issue. Adebiyi, A. A., A. O. Falase, and Y. A. Akenova. 1999. "Left Ventricular The prevention and treatment of CVDs has become a Systolic Function of Nigerians with Sickle Cell Anaemia." Annals of priority for some multilateral agencies. The WHO is active Oncology 24 (98): 27­32. in promoting prevention strategies, including tobacco con- Akinkugbe, O. O. 1976. "Epidemiology of Hypertension and Stroke in Africa." Mongor Citation 29: 28­42. trol and hypertension treatment. The World Bank has funded Alebiosu, C. O., B. Odusan, O. B. Familoni, and A. E. A. Jaiyesimi. 2004. cardiovascular control initiatives in a few countries through "Pattern of Occurrence of Microalbuminuria among Dippers and its loans program. However, among other lenders and Non-Dippers (Essential Hypertensives) in a Nigerian University Teaching Hospital." Cardiovascular Journal of South Africa 15: 9­12. donors of funds for health care, such as the Asian AMMP (Adult Mortality and Morbidity Project). 1997. Policy Implications Development Bank and most bilateral donor agencies, CVD of Adult Mortality and Morbidity. End of Phase I Report. Dar es Salaam: remains a low priority (http://www.ichealth.org). Department for International Development. Amoah, A. G. B. 2000."Spectrum of Cardiovascular Disorders in a National Referral Centre, Ghana." East African Medical Journal 77: 648­53. Anabwani, G. M., and P. Bonhoeffer. 1996. "Prevalence of Heart CONCLUSION Disease in School Children in Rural Kenya Using Colour-Flow Echocardiography." East African Medical Journal 73: 215­17. CHD and strokes share common determinants and risk Astagneau, P., T. Lang, E. Delarocque, E. Jeannee, and G. Salem. 1992. "Arterial Hypertension in Urban Africa: An Epidemiological Study on factors and are more common in urban than rural areas. a Representative Sample of Dakar Inhabitants in Senegal." Journal of Any suggestion that people of black African descent are in Hypertension 10 (9): 1095­1101. Cardiovascular Disease | 323 Beaglehole, R., and D. Yach. 2003. "Globalisation and the Prevention and Diallo, D., B. Hveem, M. A. Mahmoud, G. Berge, B. S. Paulsen, and Control of Non-Communicable Disease: The Neglected Chronic A. Maiga. 1999. "An Ethnobotanical Survey of Herbal Drugs of Diseases of Adults." Lancet 362: 903­8. Gourma District, Mali." Pharmaceutical Biology 37 (1): 80­91. Berrios, X., T. Koponen, T. Huiguang, N. Khaltaev, P. Puska, and A. Dowse, G. K., H. Gareeboo, P. Z. Zimmet, K. G. Alberti, J. Tuomilehto, Nissinen. 1997. "Distribution and Prevalence of Major Risk Factors of D. Fareed, L. G. Brissonnette, and C. F. Finch. 1990. "High Prevalence Noncommunicable Diseases in Selected Countries: The WHO Inter- of NIDDM and Impaired Glucose Tolerance in Indian, Creole, and Health Programme." Bulletin of the World Health Organization 75: Chinese Mauritians. Mauritius Noncommunicable Disease Study 99­108. Group." Diabetes 39: 390­96. Bertrand, E. 1992. "Coronary Disease in Black Africans: Epidemiology, Drury, R. A. B. 1972. "The Mortality of Elderly Ugandans." Tropical and Risk Factors, Clinical Symptomatology and Coronary Evolution." Geographical Medicine 24: 385­92. Académie Nationale de Médecine Bulletin 176 (3): 311­23; discussion Edington, G. M. 1954."Cardiovascular Disease as a Cause of Death in Gold 323­26. Coast Africa." Transactions of the Royal Society of Tropical Medicine and Betre, M., D. Kebede, and M. Kassaye. 1997. "Modifiable Risk Factors for Hygiene 48: 419­25. Coronary Heart Disease among Young People in Addis Ababa." East Edwards, R., N. Unwin, F. Mugusi, D. Whiting, S. Rashid, J. Kissima, T. J. African Medical Journal 74: 376­81. Aspray, and K. G. Alberti. 2000. "Hypertension Prevalence and Care in Boon, N. 2003. "Cardiac Disease in HIV Infection." In Oxford Textbook of an Urban and Rural Area of Tanzania." Journal of Hypertension 18 (2): Medicine, ed. D. A. Warrell, T. M. Cox, and J. D. Firth. Oxford: Oxford 145­52. University Press. Ekra, A., and E. Bertrand. 1992. "Rheumatic Heart Disease in Africa." World Health Forum 13 (4): 331­33. Bovet, P. 1995. "The Epidemiologic Transition to Chronic Diseases in Developing Countries: Cardiovascular Mortality, Morbidity, and Risk Erasmus, R. T., T. Fakeye, O. Olukoga, A. B. Okesina, E. Ebomoyi, Factors in Seychelles (Indian Ocean). Investigators of the Seychelles M. Adeleye, and A. Arije. 1989. "Prevalence of Diabetes Mellitus in a Heart Study." Sozial- und Praventivmedizin 40 (1): 35­43. Nigerian Population." Transactions of the Royal Society of Tropical Medicine and Hygiene 83 (3): 417­18. Bovet, P., A. G. Ross, J. P. Gervasoni, M. Mkamba, D. M. Mtasiwa, C. Lengeler, D. Whiting, and F. Paccaud. 2002. "Distribution of Blood Ezenwaka, C. E., A. O. Akani, B. O. Akanji, N. C. Unwin, and C. A. Pressure, Body Mass Index and Smoking Habits in the Urban Adejuwon. 1997. "The Prevalence of Insulin Resistance and Other Population of Dar es Salaam, Tanzania, and Associations with Cardiovascular Disease Risk Factors in Healthy Elderly Southwestern Socioeconomic Status." International Journal of Epidemiology 31 (1): Nigerians." Atherosclerosis 128 (2): 201­11. 240­47. Familoni, O. B., and A. J. Ariba. 2003. "Ability of Nigerian Hypertensive Bradshaw, D., D. Bourne, M. Schneider, and R. Sayed. 1995. "Mortality Patients to Perceive Changes in Their Blood Pressure." Cardiovascular Patterns of Chronic Diseases of Lifestyle in South Africa." In Chronic Journal of South Africa 14: 195­98. Diseases of Lifestyle in South Africa, ed. J. Fourie, 5­31. Cape Town: Fauci, A. S., and H. C. Lane. 2001. "Human Immunodeficiency Virus South African Medical Research Council. (HIV) Disease: AIDS and Related Disorders." In Harrison's Principles of Bradshaw, D., P. Groenewald, R. Laubscher, N. Nannan, B. Nojilana, R. Internal Medicine, 15th ed., 1,884. New York: McGraw-Hill. Norman, D. Pieterse, and M. Schneider. 2003. Initial Burden of Disease Fisch, A., E. Pichard, T. Prazuck, H. Leblanc, Y. Sidibe, and G. Brucker. Estimates for South Africa, 2000. Cape Town: South African Medical 1987. "Prevalence and Risk Factors of Diabetes Mellitus in the Rural Research Council. Region of Mali (West Africa): A Practical Approach." Diabetologia 30: 859­62. Ceesay, M. M., M. W. Morgan, M. O. Kamanda, V. R. Willoughby, and D. R. Lisk. 1997. "Prevalence of Diabetes in Rural and Urban Populations in Gaziano, T. A., and L. Opie. 2001. "Cost-Effective Analysis of Proposed Southern Sierra Leone: A Preliminary Survey." Tropical Medicine and Hypertension Guidelines in South Africa; Beta-Blocker Therapy for International Health 2 (3): 272­77. Hypertension: Hypertension in South Africa and Efficacy of First Line Therapy." Abstract. Cardiovascular Journal of Southern Africa 12 Chinyadza, E., I. M. Moyo, T. M. Katsumbe, D. Chisvo, M. Mahari, D. E. (4): 223. Cock, and O. L. Mbengeranwa. 1993. "Alcohol Problems among Patients Attending Five Primary Health Care Clinics in Harare City." Ghanem, H., and A. H. Fredj. 1999. "Eating Habits and Cardiovascular Central African Journal of Medicine 39: 26­32. Risk Factors." Presse Medicale 28 (19): 1005­8. Cooper, R. S., C. N. Rotimi, S. Ataman, D. McGee, B. Osotimehin, S. Kadiri, Hale, L. A., V. U. Fritz, and C. J. Eales. 1998. "Do Stroke Patients Realise W. Muna, et al. 1997. "The Prevalence of Hypertension in Seven That a Consequence of Hypertension Is Stroke?" South African Medical Populations of West African Origin." American Journal of Public Health Journal 88 (4): 451­54. 87 (2): 160­68. Hutt, M. S. R. 1990. "Cancer and Cardiovascular Diseases." In Disease and Cooper, R. S., C. N. Rotimi, J. S. Kaufman, E. E. Owoaje, H. Fraser, Mortality in Sub-Saharan Africa, ed. D. T. Jamieson and R. G. Feachem. T. Forrester, R. Wilks, L. K. Riste, and J. K. Cruickshank. 1997. New York: Oxford University Press. "Prevalence of NIDDM among Populations of the African Diaspora." Hutt, M. S. R., and R. Coles. 1969. "Postmortem Findings in Hypertensive Diabetes Care 20: 343­48. Subjects in Kampala, Uganda." East African Medical Journal 46: 342­58. Cruickshank, J. K., J. C. Mbanya, R. Wilks, B. Balkau, T. Forrester, S. G. IOM (Institute of Medicine). 2005. Scaling Up Treatment for the Global Anderson, L. Mennen, A. Forhan, L. Riste, and N. McFarlane- AIDS Pandemic. Washington, DC: National Academies Press. Anderson. 2001. "Hypertension in Four African-Origin Populations: IPTT (Initiative on Pharmaceutical Technology Transfer). 2003. "A New Current `Rule of Halves,' Quality of Blood Pressure Control and Partnership for Africa's Development (NEPAD) Project for the Local Attributable Risk of Cardiovascular Disease." Journal of Hypertension Manufacture of Pharmaceuticals in Africa." http://www.iptt.net. 19: 41­46. Isezuo, A. S., A. B. O. Omotoso, A. Gaye, T. Corrah, and M. A. Araoye. 2000. Davidson, J. C. 1963. "The Incidence of Diabetes in Nyasaland." Central "One-Year Survival among Sub-Saharan Africans with Hypertensive African Journal of Medicine 9: 92­94. Heart Failure." Cardiologie Tropicale 26 (103): 57­60. Davies, J. N. P. 1948. "Endocardial Fibrosis in Africans." East African JNC (Joint National Committee). 1997. "The Sixth Report of the National Medical Journal 25: 10. Committee on Prevention, Detection, Evaluation and Treatment of 324 | Anthony Mbewu and Jean-Claude Mbanya High Blood Pressure (JNC VI)." Archives of Internal Medicine 157: Mayosi, B. M., A. D. Marais, A. D. Mbewu, and P. Byrnes. 1997."Risk Factor 2413­46. Profile of Black Patients with Acute Coronary Syndromes." Canadian Kadiri, S., and B. L. Salako. 1997. "Cardiovascular Risk Factors in Middle Journal of Cardiology 13: 242B. Aged Nigerians." East African Medical Journal 74 (5): 303­6. Mbanya, J. C., E. Minkoulou, J. Salah, and B. Balkau. 1998."The Prevalence Kapuku, G. K., G. A. Mensah, and R. S. Cooper. 1998. "Hypertension in of Hypertension in Rural and Urban Cameroon." International Journal Africa." Africa Health 20 (2): 6­8. of Epidemiology 27 (2): 181­85. Khalil, S. I., K. Gharieb, M. El-Haj, M. Khalil, and S. Hakiem. 1997. Mbanya, J. C., J. Ngogang, J. N. Salah, E. Minkoulou, and B. Balkau. 1997. "Prevalence of Congenital Heart Disease among Schoolchildren of "Prevalence of NIDDM and Impaired Glucose Tolerance in a Rural Sahafa Town, Sudan." Eastern Mediterranean Health Journal 3 (1): and an Urban Population in Cameroon." Diabetologia 40 (7): 824­29. 24­28. Mbaya, V. B. 1998. "Hypertension in East Africans and Others of African Kimbally-Kaky, G., and C. Bouramoue. 2000. "Profile and Prospects of Descent: A Review." East African Medical Journal 75: 300­03. Patients from Congo with Cardiac Insufficiency: Report of 743 Cases." Mbewu, A. D. 1998. "Can Developing Country Systems Cope with the Médecine d'Afrique Noire 47 (4): 197­203. Epidemics of Cardiovascular Disease?" Paper presented at the Heart King, H., R. E. Aubert, and W. H. Herman. 1998. "Global Burden of Health Conference, New Delhi, India. Diabetes, 1995­2025: Prevalence, Numerical Estimates, and M'Buyamba-Kabangu, J., R. Fagard, P. Lijnen, J. Staessen, M. Ditu, Projections." Diabetes Care 21: 1414­31. K. Tshiani, and A. Amery. 1986. "Epidemiological Study of Blood King, H., and P. Zimmet. 1988. "Trends in the Prevalence and Incidence Pressure and Hypertension in a Sample of Urban Bantu of Zaire." of Diabetes: Non-Insulin-Dependent Diabetes Mellitus." World Health Journal of Hypertension 4: 485­91. Statistics Quarterly 41 (3­4): 190­96. Mendez, G. F., and M. R. Cowie. 2001. "The Epidemiological Features of Kokou, O. I., A. D. Agbere, D. Y. Atakoumi, A. D. Gbadoe, E. Tsolenyanu, Heart Failure in Developing Countries: A Review of the Literature." O. Djossou-Agbodjan, S. Baeta, K. Tatagan-Agbi, and K. Assimadi. International Journal of Cardiology 80 (2­3): 213­19. 1999. "Cardiac Repercussions of Sickle Cell Anaemia in Children in Metcalf, C. A., M. N. Hoffman, K. Steyn, J. M. Katzenellenbogen, and J. Lomé: Report of 70 Cases." Archives de Pediatrie 6 (10): 1134. M. Fourie. 1996. "Design and Baseline Characteristics of a Kotto, R. M., and B. A. Bouelet. 2000. "Cardiovascular Diseases in Adults Hypertension Intervention Program in a South African Village." in Douala (Cameroon)." Cardiologie Tropicale 26 (103): 61­64. Journal of Human Hypertension 10 (1): 21­26. Kruger, H. S., C. S. Venter, and H. H. Vorster. 2001. "Obesity in African Mirabet, A. 1990. "Epidemiologic Aspects of Cerebrovascular Accidents in Women in the North West Province, South Africa Is Associated with an Tunisia. Revue Neurologique (Paris) 146 (4): 297­301. Increased Risk of Non-Communicable Diseases: The THUSA Study. Transition and Health during Urbanisation of South Africans." British Mokhobo, K. P. 1976. "Arterial Hypertension in Rural Societies." East Journal of Nutrition 86 (6): 733­40. African Medical Journal 52: 440­44. Law, C. M., P. Egger, O. Dada, H. Delgado, E. Kylberg, P. Lavin, G. H. Tang, Mollentze, W. F., A. J. Moore, A. F. Steyn, G. Joubert, K. Steyn, G. M. H. Hertzen, A. W. Shiell, and D. J. Barker. 2001. "Body Size at Birth and Oosthuizen, and D. J. Weich. 1995. "Coronary Heart Disease Risk Blood Pressure among Children in Developing Countries." Factors in a Rural and Urban Orange Free State Black Population." International Journal of Epidemiology 30 (1): 52­59. South African Medical Journal 85 (2): 90­86. Law, M. 1998. "Lipids and Cardiovascular Disease." In Evidence Based Msamati, B. C., and P. S. Igbigbi. 2000. "Anthropometric Profile of Urban Cardiology, ed. S. Yusuf, J. A. Cairns, A. J. Camm, E. L. Falen, and B. J. Adult Black Malawians." East African Medical Journal 77: 364­68. Gersh, 191­205. London: British Medical Journal Books. Muna, W. F. 1993a. "The Importance of Cardiovascular Research in Africa Lawlor, D. A., G. Davey Smith, and S. Ebrahim. 2002. "Birth Weight of Today." Ethnicity and Disease 3 (Suppl.): S8­12. Offspring and Insulin Resistance in Late Adulthood: Cross Sectional ------. 1993b. "Cardiovascular Disorders in Africa." World Health Survey." British Medical Journal 325 (7360): 359. Statistics Quarterly 46 (2): 125­33. Levitt, N. S., J. M. Katzenellenbogen, D. Bradshaw, M. N. Hoffman, and Murray, C. J., and A. Lopez. 1996. The Global Burden of Disease. F. Bonnici. 1993. "The Prevalence and Identification of Risk Factors for Washington, DC: World Bank. NIDDM in Urban Africans in Cape Town, South Africa." Diabetes Care ------. 1997. "Mortality by Cause for Eight Regions of the World: Global 16: 601­7. Burden of Disease Study." Lancet 349: 1269­76. Lodenyo, H. A., S. O. McLigeyo, and E. N. Ogola. 1997. "Cardiovascular Disease in Elderly In-Patients at the Kenyatta National Hospital, Nan, L., J. Tuomilehto, G. Dowse, P. Zimmet, H. Gareeboo, P. Chitson, and Nairobi, Kenya." East African Medical Journal 74: 647­51. H. J. Korhonen. 1991. "Prevalence and Medical Care of Hypertension in Four Ethnic Groups in the Newly-Industrialized Nation of Lore, W. 1993. "Epidemiology of Cardiovascular Diseases in Africa with Mauritius." Journal of Hypertension 9 (9): 859­66. Special Reference to Kenya: An Overview." East African Medical Journal 70: 357­61. Nethononda, M. R., M. R. Essop, A. D. Mbewu, and J. S. Galpin. 2004. "Coronary Artery Disease and Risk Factors in Black South Africans-- Magula, N. P., and B. M. Mayosi. 2003. "Cardiac Involvement in HIV- A Comparative Study." Ethnicity and Disease 14 (4): 515­19. Infected People Living in Africa: A Review." Cardiovascular Journal of South Africa 14: 231­37. Njelekela, M., H. Negishi, Y. Nara, M. Tomohiro, S. Kuga, T. Noguchi, T. Kanda, et al. 2001. "Cardiovascular Risk Factors in Tanzania: A Martorell, R., L. K. Khan, M. L. Hughes, and L. M. Grummer-Strawn. 2000. Revisit." Acta Tropica 79 (3): 231­39. "Obesity in Women from Developing Countries." European Journal of Clinical Nutrition 54 (3): 247­52. Nkoua, J. L., B. Tsombou, and C. Bouramoue. 1999. "Non Rheumatic Pericarditis with Effusion: Causes, Outcome and Relation to HIV- Mayanja-Kizza, H., E. Gerwing, M. Rutakingirwa, R. Mugerwa, and Infection." Annals of Oncology 25 (97): 3­6. J. Freers. 2000. "Tropical Endomyocardial Fibrosis in Uganda: The Tribal and Geographic Distribution, and the Association with Ogunnowo, P. O., W. O. Odesanmi, and J. J. Andy. 1986. "Coronary Artery Eosinophilia." Cardiologie Tropicale 26 (103): 45­48. Pathology of 111 Consecutive Nigerians." Transactions of the Royal Matenga, J. 1997. "Stroke Incidence Rates among Black Residents of Society of Tropical Medicine and Hygiene 80 (6): 923­26. Harare--A Prospective Community-Based Study." South African Ohwovoriole, A. E., J. A. Kuti, and S. I. Kabiawu. 1988. "Casual Blood Medical Journal 87 (5): 606­9. Glucose Levels and Prevalence of Undiscovered Diabetes Mellitus in Cardiovascular Disease | 325 Lagos Metropolis Nigerians." Diabetes Research and Clinical Practice ------. 2004."Chronic Diseases." In South African Health Review 2003/04, 4 (2): 153­58. 175­90. Durban: Health Systems Trust. Okesina, A. B., D. P. Oparinde, K. A. Akindoyin, and R. T. Erasmus. 1999. Rosman, K. D. 1986. "The Epidemiology of Stroke in an Urban Black "Prevalence of Some Risk Factors of Coronary Heart Disease in a Rural Population." Stroke 17 (4): 667­69. Nigerian Population." East African Medical Journal 76: 212­16. Sarti, C., D. Rastenyte, Z. Cepaitis, and J. Tuomilehto. 2000. "International Omar, M. A., M. A, Seedat, A. A. Motala, R. B. Dyer, and P. Becker. 1993. Trends in Mortality from Stroke, 1968 to 1994." Stroke 31 (7): 1588­601. "The Prevalence of Diabetes Mellitus and Impaired Glucose Tolerance Seedat, Y. K. 1998. The Prevalence of Hypertension and the Status of in a Group of Urban South African Blacks." South African Medical Cardiovascular Health in South Africa. Ethnicity and Disease 8 (3): Journal 83 (9): 641­43. 394­97. Omotoso, A., and A. Kane. 2000. "Intraventricular Conduction Block in ------. 2000. Hypertension in Developing Nations in Sub-Saharan Adult Nigerians with Hypertensive Heart Disease; Epidemiological Africa. Journal of Human Hypertension 14 (10­11): 739­47. Study of Cardiovascular Disease and Risk Factors in Senegal: Seedat, Y. K., M. A. Seedat, and D. B. Hackland. 1982. "Prevalence of Cardiology--Out of Africa." Cardiovascular Journal of Southern Africa Hypertension in the Urban and Rural Zulu." Journal of Epidemiology 11 (3): 173­74. and Community Health 36: 256­61. Omran, A. R. 1971. "The Epidemiological Transition: A Theory of the Sen, K., and R. Bonita. 2000. "Global Health Status: Two Steps Forward, Epidemiology of Population Change." Millbank Memorial Fund One Step Back." Lancet 356: 577­82. Quarterly 49: 509­38. Sever, P. S., D. Gordon, W. S. Peart, and P. Beighton. 1980. "Blood Pressure ------. 1983. "The Epidemiologic Transition Theory. A Preliminary and Its Correlates in Urban and Tribal Africa." Lancet 2 (8185): 60­64. Update." Journal of Tropical Pediatrics 29 (6): 305­16. Shaper, A. G. 1974. "Communities without Hypertension." In Opie, L., and K. Steyn. 1995. "Rationale for the Hypertension Guidelines Cardiovascular Disease in the Tropics, ed. A. G. Shaper, M. S. R. Hutt, for Primary Care in South Africa." South African Medical Journal 85 and Z. Fejfar, 77­83. London: British Medical Association. (12): 1325­28. Sliwa, K., A. Damasceno, and B. M. Mayosi. 2005. "Epidemiology and Osuntokun, B. O., A. O. Adeuja, B. S. Schoenberg, O. Bademosi, V. A. Etiology of Cardiomyopathy in Africa." Circulation 112: 3577­83. Nottidge, A. O. Olumide, O. Ige, F. Yaria, and C. L. Bolis. 1987. "Neurological Disorders in Nigerian Africans: A Community-Based Sobngwi, E., J. C. Mbanya, N. C. Unwin, A. P. Kengne, L. Fezeu, E. M. Study." Acta Neurologica Scandinavica 75 (1): 13­21. Minkoulou, T. J. Aspray, and K. G. Alberti. 2002. "Physical Activity and Its Relationship with Obesity, Hypertension and Diabetes in Urban Owoaje, E. E., C. N. Rotimi, J. S. Kaufman, J. Tracy, and R. S. Cooper. 1997. and Rural Cameroon." International Journal of Obesity Related "Prevalence of Adult Diabetes in Ibadan, Nigeria." East African Medical Metabolic Disorders 26 (7): 1009­16. Journal 74: 299­302. Statistics South Africa. 2001. Recorded Deaths, 1996. Report 03-09-01 Oyoo, G. O., and E. N. Ogola. 1999. "Clinical and Socio Demographic (1996). Pretoria: Statistics South Africa. Aspects of Congestive Heart Failure Patients at Kenyatta National Steenkamp, J. H., I. W. Simson, and W. Theron. 1992. "Cardiovascular Hospital, Nairobi." East African Medical Journal 76: 23­27. Causes of Death at Tshepong Hospital in 1 year, 1989­1990. A Peters, V. J. 1983. "Developing In-House Audiovisual Programs For Use In Necropsy Study." South African Medical Journal 81 (3): 142­46. Diabetes Education." Diabetes Education 9 (2): 11­18. Steinberg, W. J., D. L. Balfe, and H. G. Kustner. 1988. "Decline in the Politzer, W. M., and S. B. Sachs. 1967. "Incidence of Diabetes Mellitus in Ischaemic Heart Disease Mortality Rates of South Africans, the Peri-Urban Bantu: Antenatal Surveys." South African Medical 1968­1985." South African Medical Journal 74 (11): 547­50. Journal 41 (14): 359­60. Steyn, K. 1998. South Africa Demographic and Health Survey 1998. Pretoria: Poulter, N. R. 1988. "Longitudinal Study of BP among Rural/Urban National Department of Health, chaps. 11­13. Immigrants in Kenya." In Ethnic Factors in Health and Disease, ed. D. K. Steyn, K., and J. Fourie, eds. 1991. BRISK Study Methodology: Coronary Cruikshank and D. G. Beevers. Bristol: IOP Publishing. Heart Disease Risk Factor Study in the African Population of the Cape Poulter, N. R., K. T. Khaw, B. E. C. Hopwood, M. Mugambi, W. S. Peart, Peninsula. MRC Technical Report 1. Cape Town: South African G. Rose, and P. S. Sever. 1984. "Blood Pressure and Associated Factors Medical Research Council. in a Rural Kenyan Community." Hypertension 6: 810­13. Steyn, K., J. Fourie, C. Lombard, J. Katzenellenbogen, L. Bourne, and Pretoria Department of Health, Medical Research Council, and Measure P. Jooste. 1996. "Hypertension in the Black Community of the Cape DHS+. 2002. South African Demographic and Health Survey 1998. Full Peninsula, South Africa." East African Medical Journal 73: 758­63. Report. Pretoria: Department of Health. http://www.doh.gov.za/facts/ Steyn, K., J. E. Rossouw, P. L. Jooste, D. O. Chalton, E. R. Jordaan, P. C. J. 1998/sadhs98. Jordaan, M. Steyn, and A. S. P. Swanepoel. 1993. "The Intervention Ranjith, N., R. J. Pegoraro, L. Rom, P. A. Lanning, and D. P. Naidoo. 2004. Effects of a Community-Based Hypertension Control Programme in "Renin-Angiotensin System and Associated Gene Polymorphisms in Two Rural South African Towns: The CORIS Study." South African Myocardial Infarction in Young South African Indians." Cardiovascular Medical Journal 83 (12): 885­91. Journal of South Africa 15: 22­26. Swai, A. B., D. G. McLarty, H. M. Kitange, P. M. Kikima, S. Tatalla, N. Keen, Razum, O. 1996. "Monitoring Cardiovascular Disease in Zimbabwe: L. M. Chuwa, and K. G. Alberti. 1993. "Low Prevalence of Risk Factors A Review of Needs and Options." Central African Journal of Medicine for Coronary Heart Disease in Rural Tanzania." International Journal of 42: 120­24. Epidemiology 22 (4): 651­59. Reddy, K. S., and S. Yusuf. 1998. "Emerging Epidemic of Cardiovascular Teuscher, T., P. Baillod, J. B. Rosman, and A. Teuscher. 1987. "Absence of Disease in Developing Countries." Circulation 97: 596­601. Diabetes in a Rural West African Population with a High Reddy, P. 1997. "Selling Deception and Disease: Tobacco Control." Paper Carbohydrate/Cassava Diet." Lancet 1 (8536): 765­68. presented at the Fourth International Conference on Preventive Thiam, M., G. Cloatre, F. Fall, X. Theobald, and J. L. Perret. 2000. Cardiology, Montreal, June­July 1997. Abstract. Canadian Journal of "Ischaemic Cardiopathy in Africa: Experience at the Hôpital Principal Cardiology. in Dakar." Médecine d'Afrique Noire 47 (6): 281­84. 326 | Anthony Mbewu and Jean-Claude Mbanya van der Sande, M. A., S. M. Ceesay, P. J. Milligan, O. A. Nyan, W. A. Banya, Walker, R. W., M. Rolfe, P. J. Kelly, M. M. George, and O. F. James. 2003. A. Prentice, K. P. McAdam, and G. E. Walraven. 2001. "Obesity and "Mortality and Recovery after Stroke in The Gambia." Stroke 34 (7): Undernutrition and Cardiovascular Risk Factors in Rural and Urban 1604­9. Gambian Communities." American Journal of Public Health 91 (10): WHO (World Health Organization). 1999. The World Health Report 1641­44. 1999--Making a Difference. Geneva: WHO. Vaughan, J. P. 1977. "A Brief Review of Cardiovascular Disease in Africa." ------. 2002. The World Health Report 2002--Reducing Risks, Promoting Transactions of the Royal Society of Tropical Medicine and Hygiene 71: Healthy Life. Geneva: WHO. 226­31. ------. 2004. The World Health Report 2004--Changing History. Geneva: Vorster, H. H. 2002. "The Emergence of Cardiovascular Disease during WHO. Urbanisation of Africans." Public Health and Nutrition 5 (1A): 239­43. Wiredu, E. K., and P. K. Nyame. 2001. "Stroke-Related Mortality at Korle Walker, A. R., and P. Sareli. 1997. "Coronary Heart Disease: Outlook for Bu Teaching Hospital, Accra, Ghana." East African Medical Journal 78: Africa." Journal of the Royal Society of Medicine 90 (1): 23­27. 180­84. Walker, R. W., D. G. McLarty, H. M. Kitange, D. Whiting, G. G. Masuki, Yusuf, S., S. Hawken, S. Ounpuu, T. Dans, A. Avezum, F. Lanas, D. M. Mtasiwa, H. Machibya, N. Unwin, and K. G. Alberti. 2000. M. McQueen, et al. 2004. "Effect of Potentially Modifiable Risk Factors "Stroke Mortality in Urban and Rural Tanzania. Adult Morbidity and Associated with Myocardial Infarction in 52 Countries (the INTER- Mortality Project." Lancet 355: 1684­87. HEART Study)." Lancet 364: 937­52. Cardiovascular Disease | 327 Chapter 22 Mental Health and the Abuse of Alcohol and Controlled Substances Florence K. Baingana, Atalay Alem, and Rachel Jenkins Mental disorders include depression, anxiety, schizophrenia, THE SCOPE OF MENTAL DISORDERS and psychosocial and mental disorders as consequences of alcohol and substance abuse, conflicts and complex Mental disorders are increasingly prevalent in Sub-Saharan emergencies, the human immunodeficiency virus/acquired Africa, the consequence of persistent poverty-driven condi- immune deficiency syndrome (HIV/AIDS), and gender- tions, such as malnutrition, malaria, and AIDS; the demo- based violence; in addition, there are mental disorders of graphic transition; and the persistent conflicts prevalent children. Standardized research instruments have been in the region. The leading mental disorders, anxiety and tested and widely used in Sub-Saharan Africa, and consid- depression, are often grouped together and referred to as erable attention has been paid to the transcultural per- common mental disorders (CMD). formance of these instruments. However, challenges to Many of the disorders, such as depression and anxiety, epidemiological research still exist, including unreliable are potentially preventable or treatable with currently avail- health facility records, noninclusion of mental disorders able interventions. Increasingly, epidemiological studies in the health management information systems (HMISs), carried out in Sub-Saharan Africa show that some mental and lack of any disease surveillance system that includes disorders are more prevalent there than in other areas of the mental disorders. world, and some, such as depression resulting from the con- This chapter discusses measurement and data sources, sequences of conflicts and HIV/AIDS, especially among the burden of mental disorders and alcohol abuse in Sub- orphans and other vulnerable children, are overrepresented Saharan Africa, current interventions, and effective treat- in Sub-Saharan Africa. ment strategies relevant to the context of the region. It Cultural and religious issues influence the value placed by concludes with recommendations for research as well as society on mental health, the presentation of symptoms, implications for policy formulation and development. An illness behavior, access to services, pathways through care, appendix supplies selected data on mental health disorders. the way individuals and families manage illness, the way the 329 community responds to illness, the degree of acceptance and Considerable attention has been given to the transcultural support experienced by the person with the illness, and the performance of several diagnostic tools. Goldberg and col- degree of stigma and discrimination experienced by that leagues (1997) found that the widely used screening instru- person. Therefore, cultural and other contextual issues are ment the General Health Questionnaire (GHQ) performed important considerations in developing locally appropriate just as well in detecting cases of depression and anxiety in mental health policy and programs. low-income countries as in the Western world. Bolton (2001a), Wilk and Bolton (2002), and Bolton, Neugebauer, and Ndogoni (2002) investigated how people in an African MEASUREMENT AND DATA SOURCES community severely affected by HIV view the mental health effects of the epidemic, and they used the data to investigate Although community surveys date from nearly 100 years the validity of Western concepts of depression and posttrau- ago, it is only in the last four decades that they have provid- matic stress disorder (PTSD) in the rural Ugandan commu- ed adequate diagnostic information based on standardized nity studied. Ethnographic methods (those that take into the methods of assessment, allowing the comparison of research field certain developed viewpoints and techniques but also from different locations and from different levels. acknowledge that because individual cultures are unique A useful framework, originally devised for understanding they can be evaluated only according to their own values and the pathway by which individuals become defined as standards, wary of the ethnocentric belief that one's own mentally ill--that is, passing through primary care and culture is superior in every way to all others) were used, and eventually reaching specialist mental health services--is the the participants were able to describe two, independent, organization of epidemiological data into groupings. These depression-like syndromes resulting from the HIV epidemic. groupings are defined by how far along the pathway the No syndromes similar to posttraumatic stress syndrome population under study has come, in order to ensure that were found. The authors concluded that people recognize like is compared with like (Goldberg and Huxley 1992). The depression syndromes and consider them consequences of framework comprises five levels: (a) all individuals in the the HIV epidemic. Bolton and colleagues (2003) evaluated community, (b) all individuals attending at the primary the feasibility of conducting controlled studies in Africa and care level, (c) those in primary care who have been diag- found the controlled trial to be feasible in the local setting. nosed by their doctor or nurse, (d) all those referred to spe- Kaaya and colleagues (2002) carried out a study to test the cialist services, and (e) those who have been admitted to a validity of the Hopkins Symptom Checklist-25 (HSCL-25) hospital. The framework postulates a set of four filters among HIV-positive women in Tanzania. They found the between the five levels; the filters are influenced by the ill- internal consistency of the HSCL-25 and the HSCL-15 to be ness behavior of the patient, access to primary health care adequate. The HSCL-25 demonstrated its usefulness as a services, the ability of the primary care team to detect disor- screening tool for depression. der, and the team's capacity to refer patients to higher or Epidemiological studies have been carried out in Kenya lower levels of care, and access to specialist services and in- (Kiima et al. 2004) using the UK Clinical Interview Schedule patient care. (revised), and in Burundi, a survey of national welfare indi- Standardized research instruments, both questionnaires cators was carried out (Baingana et al. 2004) that included and structured or semistructured interviews, have been the 12-item GHQ and the 5-item Alcohol Use Disorders developed, tested, and used widely around the world over the Identification Test (AUDIT). The Kenya mental health survey last few decades, allowing estimates of prevalence, incidence, also used the Core Welfare Indicators Questionnaire (CWIQ), outcome, and examination of associated risk factors (for an instrument developed by the World Bank for rapid pop- example, Thompson Psychiatric Research Methods, Hopkins ulation-level assessments of welfare indicators (World Bank Symptom Checklist, General Health Questionnaire, Beck's 2001), and the AUDIT. In 1998, 12 psychological questions Depression Inventory, Harvard Trauma Questionnaire). adapted from the GHQ-12 were integrated into the Burundi Whereas studies in the 1950s and 1960s showed that the Household and Living Standards Survey. Analysis of the reliability of psychiatric diagnoses was often low, the intro- data revealed two indicators of distress, similar to those duction of international diagnostic systems with guides, found by Goldberg and colleagues (1997) while using the structured interviews, and operational definitions has trans- GHQ-12. The two indicators were internally validated formed the situation. (Baingana et al. 2004). 330 | Florence K. Baingana, Atalay Alem, and Rachel Jenkins Strauss and colleagues (1995), in assessing the predictive large number of the studies use hospital-based data, yet value of a screening questionnaire for depression and anxi- many of the patients do not have access to or knowledge of ety, found that general practitioners could correctly diag- the mental health services available in their vicinity. nose depression in 3.2 percent of patients. The screening Although prevalence studies for mental disorders in questionnaire had a 42 percent chance of correctly iden- Sub-Saharan Africa are scarce, studies from the rest of the tifying depression and a 97 percent chance of correctly world can be extrapolated to the Sub-Saharan region. identifying patients who did not have depression. The study Epidemiological studies of communities (for example, confirmed the low identification rate for depression among Jenkins 1998; Kessler et al. 1994; Kessler et al. 2005), of peo- general practitioners, highlighting how unreliable patient ple at work (Jenkins 1985), and of people in primary health records are as sources for epidemiological studies. care (Demyttenaere et al. 2004; Ormel et al. 1994; Sartorius South Africa's National Non-Natural Mortality Surveil- et al. 1996; Ustun et al. 2004) from all regions of the world lance System, which began operation in 1998, is an excellent have shown that depression and anxiety are common every- source for suicide mortality. "Non-natural mortality" refers where. They contribute to sickness absence and labor to deaths from homicide, suicide, accidents, and undeter- turnover (Jenkins 1985) and form a significant contribution mined causes. An evaluation carried out in 2001 found sen- to the overall public health burden (Murray and Lopez sitivity to range from 65 to 95 percent for manner of death, 1996). In 1996, major unipolar depression was projected to with a positive predictive value that ranged from 74 to be number two as a leading cause of disability in 2020 80 percent for manner of death and 71 to 82 percent for (Murray and Lopez 1997a); estimates of the Global Burden mechanism of death. Maintenance costs are estimated to be of Disease carried out in 2000 found depression to be the R 8.00 (US$1.00) per case registered. leading cause of disability, accounting for 12 percent of the However, there still exist challenges to epidemiological global disability burden (Ustun et al. 2004). research for mental disorders in Sub-Saharan Africa. Health facility records are not reliable, mental disorders are not Global Burden of Disease 2000 Study included as separate items in the HMIS, and very few cross- sectional or longitudinal studies have been carried out. No The Global Burden of Disease was launched by the World disease surveillance system includes mental disorders, and Health Organization (WHO) in the 1990s, and the first set no censuses, registries, or other administrative data include of data was published in 1996 (Murray and Lopez 1996). In mental disorders. 2000 WHO carried out the second assessment of the Global A PubMed search was carried out with the key words Burden of Disease, the GBD 2000 (Mathers et al. 2002). This "mental disorders," "depression," "suicide," and "Sub- is the most up-to-date GBD data available. The GBD 2000 Saharan Africa." Few epidemiological studies carried out in study estimated the incidence and prevalence and the mor- Sub-Saharan Africa have been published in peer-reviewed tality rates of major depression, bipolar disorder, schizo- journals. The best data sources are from South Africa and phrenia, panic disorder, and obsessive-compulsive disorder Nigeria and a few from Kenya and Zambia. Additional data in the Sub-Saharan Africa WHO subregions (table 22.1). are potentially available in unpublished surveys, particularly The study also estimated years lived with disability (YLDs) in Francophone Africa. and disability-adjusted life years (DALYs) for these disorders (table 22.2). EPIDEMIOLOGY Unipolar Depression. In the GBD 1996, unipolar depres- sion, which differs from bipolar depressive disorder in that Major challenges to epidemiological research are limited it presents with recurring depressive episodes without any capacity in the use of international classification systems for manic episodes, unlike bipolar disorder, which has both coding of disorders, noninclusion of mental and neurologi- manic and depressive episodes, was estimated to be the lead- cal disorders as separate categories in the HMIS, nonstan- ing cause of the nonfatal disease burden in the world in dardized instruments for epidemiological research that have 1990, accounting for 10.7 percent of total YLD. It was the not been validated for use in all areas of the subcontinent, fourth leading cause of total disease burden, accounting for and limited capacity and resources to carry out comprehen- 3.7 percent of total DALYs (Murray and Lopez 1997b). In sive and scientifically sound community assessments. A the GBD 2000 study, unipolar depression remains the Mental Health and the Abuse of Alcohol and Controlled Substances | 331 Table 22.1 Age-Standardized Incidence, Prevalence, and Mortality Rate Estimates for WHO AFR Epidemiological Subregions, 2000 (per 100,000 people) Incidence Prevalence Mortality Disorder/Area Males Females Males Females Males Females Schizophrenia AFR D 18 21 343 378 0 0 AFR E 20 24 349 418 1 0 World 19 20 422 423 0 0 PTSD AFR D 45 121 216 552 -- -- AFR E 44 126 212 558 -- -- World 44 121 208 559 -- -- Panic disorder AFR D 32 63 309 613 -- -- AFR E 32 63 309 613 -- -- World 30 61 319 631 -- -- Obsessive­compulsive disorder AFR D 77 83 586 790 -- -- AFR E 77 83 586 790 -- -- World 58 77 376 522 -- -- Unipolar depressive disorder AFR D 2,851 4,345 1,426 2,173 319 621 AFR E 2,851 4,345 1,426 2,173 319 621 World 3,199 4,930 1,607 2,552 323 630 Bipolar disorders AFR D 26 25 482 450 0 0 AFR E 26 25 482 450 0 0 World 26 25 467 472 0 0 Source: Mathers et al. 2002. Note: -- not available. leading cause of YLD, accounting for 11.9 percent of total prevalence of 3.7 percent for males and 15.3 percent for YLD, and also remains the fourth leading cause of total dis- females in a population age 19 to 93 years. This prevalence ease burden, accounting for 4.4 percent of total DALYs was substantially higher than that found in other regions of (Mathers et al. 2002). In Butajira, Ethiopia, a demographic the world (Mathers et al. 2002). study site for the University of Addis Ababa and the ministry of health of Ethiopia, it was found that, using the DALY Bipolar Disorder. Bipolar disorder is a chronic disease method, depression contributed 7 percent to the total with periods of depression and elevated mood and with disease burden (Abdulahi, Mariam, and Kebede 2001).1 remissions and relapses between them. In the 1990 GBD, it was estimated to be the seventh leading cause of the nonfa- Panic Disorder. In the GBD 1996, panic disorder was esti- tal burden of disease, accounting for 3 percent of the total mated to be the 27th leading cause of the nonfatal burden of YLD (Murray and Lopez 1996). In the GBD 2000, bipolar disease in the world (Murray and Lopez 1996). In the GBD disorder accounts for 2.5 percent of total YLD (Mathers 2000 the estimated burden of panic disorder increased et al. 2002). slightly, accounting for 1.2 percent of the total YLD. One of the data sources used for this estimate was a study carried Schizophrenia. The seventh leading cause of YLD at the out in Lesotho, which found an estimate of a one-month global level, schizophrenia accounts for 2.8 percent of total 332 | Florence K. Baingana, Atalay Alem, and Rachel Jenkins Table 22.2 YLD, YLL, and DALY Estimates for WHO AFR Epidemiological Subregions, 2000 YLD per 100,000 YLL per 100,000 YLD YLL DALYs Disorder/Area Males Females Males Females (000s) (000s) (000s) Schizophrenia AFR D 250 246 2 1 828 6 834 AFR E 237 249 3 1 820 7 827 World 259 252 5 4 15,427 263 15,690 PTSD AFR D 27 68 -- -- 160 0 160 AFR E 25 68 -- -- 158 0 158 World 29 78 -- -- 3,230 0 3,230 Panic disorder AFR D 77 152 -- -- 382 0 382 AFR E 77 151 -- -- 386 0 386 World 74 145 -- -- 6,591 0 6,591 Obsessive­compulsive disorder AFR D 107 144 -- -- 420 -- 420 AFR E 107 146 -- -- 428 -- 428 World 67 90 -- -- 4,761 -- 4,761 Unipolar depressive disorder AFR D 514 786 0 0 2,172 0 2,172 AFR E 507 768 0 0 2,154 0 2,154 World 851 1,302 0 0 64,963 0 64,963 Bipolar disorders AFR D 261 245 -- -- 845 -- 845 AFR E 261 244 -- -- 852 -- 852 World 226 224 -- -- 13,610 36 13,645 Source: Mathers et al. 2002. Note: -- not available; YLL years of life lost. global YLD in the GBD 2000 study, up from 10th place about the same percentage as schizophrenia (Murray (2.6 percent of YLD) in 1990 (Mathers et al. 2002; Murray and Lopez 1996). In the GBD 2000, PTSD has increased to and Lopez 1996). In an Ethiopian Burden of Disease study, 0.6 percent of total YLD (Mathers et al. 2002). 4 percent of the total disease burden was due to schizophre- nia (Abdulahi, Mariam, and Kebede 2001). Sartorius and Obsessive-Compulsive Disorder. Obsessive-compulsive colleagues (1986) estimate an incidence rate for schizophre- disorder was estimated to be the 11th cause of the nonfatal nia of 10 cases per 10,000 people. Mathers and colleagues burden of disease in the world, accounting for 2.2 percent of (2002) estimate the incidence rate for Sub-Saharan Africa at total YLD in 1990 (Murray and Lopez 1996). In the GBD 0.2 for males and 0.3 for females per 1,000 people, age- 2000, obsessive-compulsive disorder now accounts for adjusted to between 0.4 and 0.53 percent. In a large semi- 2.5 percent of total YLD (Mathers et al. 2002). urban and rural population study in Ethiopia the prevalence of schizophrenia was found to be 4.7 per 1,000 people Data Specific to Sub-Saharan Africa (Kebede et al. 2003). Although data specific to Sub-Saharan Africa are more dif- Posttraumatic Stress Disorder. In the GBD 1990, PTSD ficult to find, the number of epidemiological mental was estimated to account for 0.4 percent of the total YLD, health studies being carried out in the last decade has Mental Health and the Abuse of Alcohol and Controlled Substances | 333 increased. This is due to an increase in the number of with medically unexplained physical symptoms as well as African universities that are training psychiatrists, an the symptoms of depression, and anxiety (Patel et al. 1995). increase in the number of African psychiatrists who have People with CMDs are frequent attenders at general primary been trained in research methods and who are practicing in health care clinics and medical and surgical inpatient beds, Africa, an increase in the number of African universities because they are specifically seeking help for their disorder, partnering with Western and northern universities, as well because they have a coexisting physical illness, or because as an increasing number of African institutions participat- they are somatizing their mental disorder. CMDs are signif- ing in multi-site studies led by WHO or by U.S. research icant contributors to the workload in general health clinics. institutes. There is also a wealth of information coming out of South Africa, which had a much stronger research culture Schizophrenia. Schizophrenia is found in all countries and but was closed off to the rest of Africa in the apartheid years. cultures and has a lifetime prevalence of between 7 and 9 per 1,000 people (Jablensky et al. 1992). The point prevalence Depression. There have been at least three geographically varies between about 2 and 5 per 1,000. Collaborative stud- localized but well-conducted epidemiological surveys of ies by the WHO have shown that the prevalence of schizo- depression in Sub-Saharan Africa. Hollifield and colleagues phrenia, when assessed in comparable ways, is similar in (1990) used a two-stage approach in a rural village in different countries (Jablensky et al. 1992). Although much Lesotho and found a prevalence of depression of 9 percent rarer than depression and anxiety, it too forms a significant when alcohol abuse was corrected for, assuming that alcohol contribution to the overall public health burden because of use preceded depressive disorder. The researchers noted that the chronicity, deterioration, and extreme social disability in 19 percent of subjects with panic and depressive disorder a significant proportion of sufferers. The outcome of schiz- had comorbid alcohol use. Rumble and colleagues (1996) ophrenia is more favorable in developing countries than in note a similar high prevalence of depression in people in the West, but the reasons are not yet clear (IOM 2001; a rural South African village, in a study that screened for Jablensky et al. 1992; Leff et al. 1992; Sartorius et al. 1986). symptoms using an adapted version of the Self Report Jablensky recently reviewed the epidemiology of schizo- Questionnaire­25-item version (SRQ-25) and the Present phrenia (IOM 2001) and concluded that there have been few State Examination (PSE) as a second-stage instrument. systematic surveys of psychoses in Africa, although there The weighted prevalence for unipolar depressive disorder are plenty of service-based clinical studies. An exception is using the CATEGO (a computerized classification system) a recent, well-designed community survey in an area of was 18 percent. More than half of the psychiatric morbidity Ethiopia with a population of 100,000 in the age range 15 to detected was attributed to depressive disorders. In a home- 49 years. He concluded that the reported point prevalence stead survey in two villages in Uganda, where the Luganda of schizophrenia in most areas of the developing world version of the PSE was used, Orley, Blitt, and Wing (1979) where epidemiological surveys have been conducted is found that 14.3 percent of males and 22.6 percent of females comparable with that in the developed world. Taking into among adults age 18 to 65 met criteria for depressive disor- account such factors as the higher mortality among people der using the CATEGO classification system. Estimated with serious mental disorders and incomplete ascertain- prevalence of depression in women in Harare, Zimbabwe, ment of a proportion of cases, it is likely that the reported was reported to be 30 percent (Abas and Broadhead 1994) rates are underestimates of the true prevalence of the and the incidence rate of depression in the same population disorder. was estimated at 18 percent (Broadhead and Abas 1994). The incidence rate of major depression was found to be Posttraumatic Stress Disorder. Forty-one percent of all 15.6 percent in a rural general practice of South Africa deaths in the WHO Sub-Saharan Africa region are from (Strauss et al. 1995) and 15.5 percent in Rwanda five years intentional injuries, the highest rates being for males age after the genocide of April 1994 (Bolton, Neugebauer, and 15 to 29 years (56 percent) and 30 to 34 years (53 percent). Ndogoni 2002). The Sub-Saharan Africa region has the highest rates of death due to war-related injuries in the world, with a rate of Common Mental Disorders. Common mental disorders is a 22 percent of all war-related injuries and 32 per 100,000 term used to discuss both anxiety and depression when they people (WHO 2002b). Estimates for psychosocial and occur in the primary health care setting, often presenting mental disorders resulting from conflicts were 15.5 percent 334 | Florence K. Baingana, Atalay Alem, and Rachel Jenkins in Rwanda five years after the genocide (Bolton 2001b; Nigeria, South Africa, and Zambia. In South Africa from Bolton, Neugebauer, and Ndogoni 2002). Studies carried 1984 to 1986 up to 0.37 percent of all admissions to hospi- out in Africa and other parts of the world indicate rates of tals (Okulate 2001) and 1.3 percent of all national mortality 20 to 60 percent for depression, anxiety symptoms, and were suicide attempts (Flisher and Parry 1994). Parasuicide PTSD among children and women (de Jong 2002; Mollica (attempted suicide) rates are much higher. In a study of et al. 1998; Mollica et al. 1999). 10,984 patients seen in a hospital in Durban, 17.7 percent were referred to the Department of Psychiatry because of Alcohol Abuse. Adult per capita alcohol consumption is parasuicide (Schlebusch 1985). In Nairobi, Kenya, suicide generally estimated by dividing the sum of alcohol produc- was the fourth most common cause of death due to injuries, tion and imports less alcohol exports by the adult popula- making up 12 percent of all injury-related deaths (Muniu tion age 15 years and older (WHO 1999b). In countries et al. 1994). where the production is mainly home brews and spirits, thus In a study of high school students in Addis Ababa in 1989 not taxable, as in most of Sub-Saharan Africa, it becomes and 1990, Kebede and Ketsela (1993) found 14.3 percent extremely difficult to get an accurate estimate of consump- reporting having attempted suicide. Kebede and Alem tion. It is now widely accepted that the proportion of the (1999) found a lifetime prevalence of 0.9 percent in Addis population drinking excessively is closely related to the aver- Ababa, and Alem et al. (1999) found a lifetime prevalence of age consumption of that population (WHO 1999b). The 3.2 percent in Butajira, Ethiopia. Suicides generally occur increase in road traffic accidents, liver cirrhosis, and pancre- between the ages of 13 and 50 years and peak at 20 to atic disease as alcohol consumption increases further vali- 29 years (WHO 1999a). Most studies reported that females dates this premise. generally have a higher frequency of suicide than males. The WHO estimates a sharp increase in per capita con- Most of the suicide and parasuicide is associated with sumption of alcohol in Sub-Saharan Africa. Five of the depression or alcohol abuse, or both. Dong and Simon 13 countries with the world's highest increase in alcohol (2001) found an increase of 320 percent in organophos- consumption from 1970­72 to 1994­96 are in Sub-Saharan phate poisoning in urban Zimbabwe for the period 1995 to Africa. Lesotho ranked 1st, with a 1,817 percent increase; 2000. Breetzke (1988) reported that in South Africa, suicide Nigeria, 5th, with a 196 percent increase; Rwanda, 10th, with was more frequent among the white population (14 per a 129 percent increase; Burkina Faso, 12th, and Sudan, 100,000) than among those of mixed race (3 per 100,000) 13th, with 116 percent and 108 percent increases, respec- and blacks (0.7 per 100,000). A study of accidental and tively (WHO 1999b). Drinking is greater among males than violent death in Tanzania in women age 16 to 45 found that females and greater among the uneducated than the suicide is as common in that country as it is in the United formally educated. In the Seychelles, male drinkers consume Kingdom (CDC 2000; Setel et al. 2000). Table 22.3 summa- eight times as much alcohol as females; among black South rizes some of the studies of suicide carried out in Sub- Africans, more than twice the men drink more regularly Saharan Africa. than women; and in Zambia, four times as many men as women drink weekly. A pattern of men drinking more fre- Psychiatric Disorders among Children. Not many data quently and to the point of intoxication is prevalent across are available on the burden of mental disorders among Sub-Saharan Africa (WHO 1999b). children. This is mainly because of the lack of validated A worrying trend is that of consumption of alcohol by testing instruments sensitive to the context of Sub-Saharan children (WHO 1999b). In Namibia, 20 percent of school- Africa; the lack of specialized personnel, such as child psy- children and 75 percent of young people not in school chiatrists and child psychologists, able to carry out the abuse alcohol on weekends. In Zimbabwe, 31 percent of assessments; and limited resources. Schier, Yecunnoamlack, those age 14 years and under report using alcohol. In and Tegegne (1989) found that 6.8 percent of 1,078 chil- Lesotho, 8.8 percent of children between the ages of 10 and dren treated on pediatric wards in Ethiopia had a neu- 14 years and 4 percent of those between 5 and 9 years cur- ropsychiatric disorder. A study carried out in Kenya rently use alcohol. found that one-third of the children referred to a psycho- logical assessment center had emotional disorders as the Suicide. Suicide is an important cause of mortality in Sub- cause for their learning difficulties (Dhadphale and Saharan Africa. Studies have mainly been carried out in Ibrahim 1984). Mental Health and the Abuse of Alcohol and Controlled Substances | 335 Table 22.3 Summaries of Selected Studies on Suicide and Parasuicide Authors and year Country and study population Study findings Oguleye, Nwaorgu, and Nigeria 35% suicidal Grandawa 2002 10-year study, 23 corrosive esophagitis 75% of suicides in the second decade of life patients Granja, Zacarias, and Mozambique 9 deaths due to alleged suicide Bergstrom 2002 Retrospective study 59% were younger than 25 years of age 27 pregnancy-related deaths Mzezewa et al. 2000 Zimbabwe 89% females Prospective study of suicidal burns Median age: 25 years; range: 13­50 years 47 patients 64% housewives Mortality 68% Alem et al. 1999 Ethiopia Lifetime suicide attempt: 3.2% of population; Cross-sectional survey of 10,468 adults of these, 63% women 15­24 years most frequent age group for suicide attempt People with mental distress and problem drinking had higher lifetime prevalence for suicide attempt Kebede and Alem 1999 Ethiopia 2.7%, prevalence of current suicide ideation 10,203 adults in Addis Ababa 0.9%, lifetime prevalence 66% under the age of 25 years Current suicidal ideation more common in men than in women Decreasing risk of suicide attempt with increasing age and educational attainment Wilson and Wormald 1995 South Africa Patient had limited schooling; unemployed 27 adults who had taken battery acid Male-female ratio of 2.4 to 1 9 had diagnosable psychiatric illness Mboussou and Gabon Higher ratio for attempted suicide among women Milebou-Aubusson 1989 39 cases of suicide at a female-male ratio of 3:1 208 attempted suicide More frequent among younger age groups Odejide et al. 1986 Nigeria 76.9% under 30 years of age 39 cases of deliberate self-harm Male-female ratio of 1.4:1 51.3% students 25.6% manual workers Cummins and Allwood 1984 South Africa 10% were suicide attempts 10­15-year-olds referred to a child Peak incidence among 13 year olds psychiatry clinic Male-female ratio of 2:1 Predisposing and antecedent causes were family stress (divorce), psychiatric illness in the patient or a family member, school problems 7% made further serious suicide attempts Eferakeya 1984 Nigeria Crude suicide rate of 7 per 100,000 87% of attempters under the age of 30 years Highest age group 15­19 years (39.4%) Female-male ratio of 1:1.2 64% of attempters were students, housewives, and the unemployed Major predisposing factor was mental illness (32%) Source: Compiled by authors. Etiology and Determinants Malnutrition. Malnutrition is prevalent in Sub-Saharan Africa; from 20 to 50 percent of all children under five years The major risk factors for mental disorders can be classified are severely malnourished. The World Health Report 2002 into genetic factors; nutritional deficiencies; infection; estimates that 32 percent and 31 percent of children in exposure to environmental toxins; prenatal, perinatal, and WHO AFR D (characterized by high child and high adult neonatal factors; poverty; and trauma. 336 | Florence K. Baingana, Atalay Alem, and Rachel Jenkins mortality) and AFR E (characterized by high child and very Life Events. Life events that lead to the threat of loss or to high adult mortality), respectively, are two standard devia- actual loss, such as the death of a family member, marital tions below the weight for age.2 Malnutrition is associated separation, maternal deprivation, or loss of employment, with mild to moderate mental retardation. Micronutrient have been shown to cluster before the onset of depressive deficiencies have also been found to be a major risk factor. episodes and to influence the course of depression in both These include iodine and zinc, with rates of 37 percent in developed and developing countries. Beck (1986) has AFR D and 62 percent in AFR E not consuming the recom- described a cognitive triad that may contribute to the onset mended dietary intake. Iodine deficiency is the single most or reoccurrence of depressive episodes by increasing the risk prevalent cause of mental retardation and brain damage, of exposure to stressful life events: negative self view, nega- 25 percent of the global burden of iodine-related deficiency tive interpretation of experience, and negative view of the disorders are contributed by AFR E (WHO 2002a). Iron future. Rates of depression increase in a variety of vulnera- deficiency is prevalent in the region with hemoglobin ble groups, including refugees, neglected ethnic minority levels of 10.6 for both AFR D and AFR E. A growing body of groups, and those exposed to war trauma (Baingana et al. evidence indicates that iron deficiency anemia in early 2004; Baingana, Bannon, and Thomas 2005; Barton and childhood is associated with reduced intelligence in mid- Mutiti 1998; de Jong 2002; Green et al. 2003; Mollica et al. childhood (WHO 2002a). A study carried out in South 1998). Depression is also postulated to be high in Sub- Africa found that permanent intellectual stunting results Saharan Africa, resulting from the prevalent violence against from chronic malnutrition of infants up to four years of age women and HIV/AIDS (Baingana, Thomas, and Comblain, (Booyens, Luitingh, and van Rensburg 1977). 2005; Bouta, Frerks, and Bannon 2004). Genetic Vulnerability. There is strong evidence that Poverty. A large body of evidence demonstrates the associ- genetic vulnerability is an important part of the cause of ation between poverty and CMD. For example, a meta- schizophrenia; a person's risk of developing the disorder analysis of five cross-sectional surveys carried out in Brazil, increases steeply with the degree of genetic relatedness Chile, India, and Zimbabwe of people who sought treatment (IOM 2001). Few risk factors have been specifically identi- in primary care and the community, examining the eco- fied or validated in developing countries, although obstetric nomic risk factors for CMD, found a consistent and signifi- complications and early brain injury due to neuroinfection, cant relation between low-income countries and risk for toxic effects, other trauma, or maternal malnutrition during CMD. Similarly, a population-based study from Indonesia gestation are likely to be involved in a greater proportion of revealed that people with less education and fewer material cases of adult schizophrenia in the developing world than in possessions than others in their community were more the developed world. likely to suffer from depression (Friedman 2004). It appears Incidence and prevalence studies from developing coun- that both absolute and relative poverty are important in the tries suggest a clustering of onset of schizophrenia in early genesis of depression (Friedman 2004). adulthood, similar to that observed in developed countries, Voices of the Poor, a three-volume publication of the World although it tends to occur at an earlier age in developing Bank that reports the findings from global focus group dis- countries. The onset tends to be earlier in males than in cussions with more than 60,000 poor people, notes feelings females. An important difference between developing and of worthlessness, hopelessness, and anxiety, as well as lack of developed countries is that in the majority of developed planning for the future, as expressions of the state of being countries males have a higher morbidity than females, poor. These are some of the core symptoms of depression. whereas in some developing countries this dominance is attenuated or inverted. Given that in many developing coun- Gender. Both community and primary care­based studies tries, women have higher mortality than men, this finding indicate that women are often affected disproportionately suggests that if adjustment for mortality could be made, the by depression in both developing and developed countries risk for schizophrenia for women in developing countries (Abas and Broadhead 1997; Bean and Moller 2002; would be even higher. Causes of such a higher risk of schizo- Broadhead and Abas 1994; Patel et al. 1999; Ustun et al. phrenia among women in developing countries may involve 2004). The multiple roles assumed by women, including both biological and psychosocial factors and requires further the bearing and rearing of children, responsibility for the research. home, caring for both healthy and ill relatives, growing food, Mental Health and the Abuse of Alcohol and Controlled Substances | 337 and earning income, can lead to increased exposure to life instruments developed by the WHO (Yegomawork et al. events, social adversity, and other environmental factors. 2003). This study showed that 59 percent of women suffered Women in both developed and developing countries also from sexual violence, and 49 percent suffered physical vio- encounter difficulties in relation to their social position, lence in their lifetime. Within the 12 months prior to the aspirations, social support networks, and domestic prob- survey, 29 percent and 44 percent, respectively, had experi- lems, which may include physical or sexual abuse. enced physical and sexual violence. Very often, intimate Postpartum depression has been identified in both devel- partners are responsible for the violence. Women who suf- oped and developing countries. The greatest risk for post- fered domestic violence reported increased lifetime mental partum depression is within the first 30 days of childbirth, health problems more often than those who did not suffer and the condition can persist for up to two years. Certain these adverse life events. The etiology of suicide is linked to a practices in some developing countries, such as isolation of woman's history of sexual abuse and domestic violence; recent mothers from family and the new infant, are disrup- alcohol abuse; stressful life events, such as unwanted preg- tive to the initial mother-infant relationship and eliminate nancy or school-related pressures; as well as mental disor- the benefits of positive social supports. These practices ders, including depression and schizophrenia. An inverse have been identified as possible contributing factors to the relation has been found between suicide and education, onset of postpartum depression. In a study carried out in older age, higher social class, and other indicators of well- Zimbabwe, a brief screening questionnaire proved effective being. Table 22.3 summarizes some of the studies carried in identifying women in the eighth month of pregnancy out on suicide in Sub-Saharan Africa. who were at higher risk of postpartum depression Weiss, Longhurst, and Mazure (1999), studying risk for (Nhiwatiwa, Patel, and Acuda 1998). Such a tool may be depression in American women, found that child sexual useful in devising preventive measures aimed at implement- abuse is associated with adult-onset depression in both men ing interventions shortly after childbirth for previously and women. A study of the patient profile of a clinic for identified high-risk individuals. child abuse and neglect in South Africa found that females Violence against women, including sexual abuse of chil- made up 80 percent of the patients and that sexual abuse dren, is linked to mental disorders of these women (90 percent) was the most common presenting complaint (Mulugeta, Kassaye, and Berhan 1998). A study carried out (de Villiers and Prentice 1996). The majority of the patients in Ethiopia found that in 5 percent of female high school were young, 55 percent being below 10 years of age and students, completed rape had occurred and another 10 per- 7 percent below 3 years. Behavior problems were recorded in cent suffered attempted rape. Social isolation (33 percent), 73 percent of cases, the commonest being school problems fear and phobia (19 percent), hopelessness (22 percent), and (21 percent), masturbation (19 percent), clingy behavior suicide attempts (6 percent) were the psychological sequelae. (12 percent), and withdrawal or depression (11.5 percent). Rates of 21.1 percent have been found for domestic violence Similar findings were reported by Berard and among patients presenting at a general practitioner's office Boermeester (1999). This study also found that more sexu- (Marais et al. 1999). ally abused patients received a diagnosis of depression than Depression resulting from violence against women is pos- was expected, and they also scored higher on depression- tulated to be high in Sub-Saharan Africa (Marais et al. 1999). rating scales. Logistic regression showed that the presence of Rates of PTSD for those with a history of domestic violence suicidal symptoms and alcohol use were independently were 35.3 percent versus 2.6 percent for those without such associated with sexual abuse. The authors concluded that a history; rates of depression were 48.2 percent versus "the associations of sexual abuse with suicidal symptoms, 11.4 percent; rates of suicide attempt were 19.0 percent ver- alcohol use, and troubled family circumstances, in the con- sus 5.8 percent; and rates of substance abuse were 9.4 per- text of high unemployment, poverty, and gang-related vio- cent versus 4.7 percent. Those with major depression were lence, indicate a strong correspondence between adverse also more likely to have attempted suicide and more likely to social conditions and psychological symptoms" (Berard and have unexplained physical symptoms and to make more Boermeester 1999, 975). visits to the general practitioner. Those with depression were also more likely to have comorbid PTSD (Marais et al. 1999). HIV/AIDS. The psychiatric sequelae of HIV/AIDS are In Butajira, Ethiopia, more than 3,000 women were numerous and have etiologies that involve neurobiological systematically selected for a domestic violence study using and psychosocial factors. These include depression, anxiety 338 | Florence K. Baingana, Atalay Alem, and Rachel Jenkins disorders, manic symptoms, and atypical psychosis. social security payments, as is the case in many developing Neuropsychiatric abnormalities were present in 41 percent countries, this situation is greatly aggravated. In addition to of patients who tested positive for HIV-1 in Zaire (Perriens this effect of poverty on the individual, the economy experi- et al. 1992), and depression was found to be higher in ences a loss because of lost production from people with symptomatic seropositive individuals than in matched mental and neurological illnesses being unable to work, seronegative individuals in the WHO Neuropsychiatric either in the short, medium, or long term, and reduced pro- AIDS study (Maj et al. 1994). Kwalombota (2002), in a study ductivity from people being ill while at work. carried out in Lusaka, Zambia, on the effect of pregnancy There are also socioeconomic costs to families, including in HIV-infected women, found 85 percent to have major the cost of supporting the dependents of people with mental depressive episodes with suicidal thoughts. Those who knew and neurological illnesses. Long-term consequences include their HIV status before becoming pregnant did not show unemployment, crime, and violence in young people whose severe depressive episodes, but those who found out while childhood problems (for example, depression, conduct dis- pregnant were liable to develop major depressive illness. order, dyslexia, and other special educational needs) were In Zaire, Boivin et al. (1995) studied the impact of a not properly addressed in childhood. mother's HIV-positive status on the well-being of her chil- dren. The researchers found that "maternal HIV infection Poor Physical Health. Poor mental and neurological health compromises the labor-intensive provision of care in the is a risk factor for many physical health problems, and emo- African milieu and undermines global cognitive develop- tional distress makes people more vulnerable to physical ill- ment in even uninfected children" (p. 13). ness. Various studies carried out in the West show that Some research has been carried out in Sub-Saharan depression increases the risk of heart disease fourfold (Lett Africa on the links between HIV/AIDS and mental and neu- et al. 2004; Nemeroff, Musselman, and Evans 1998; Zellweger rological disorders. Although much of the research was done et al. 2004), even though it controlled for other risk factors in the developed world, the data can provide information on such as smoking (Hippisley-Cox, Fielding, and Pringle the gaps in knowledge and the possible consequences of 1998). Lack of control at work is also associated with being HIV positive in Sub-Saharan Africa. Table 22.4 sum- increased risk of cardiovascular disease (Marmot et al. 1997). marizes several of the studies on HIV/AIDS in Sub-Saharan Sustained stress or trauma increases susceptibility to viral Africa. infection and physical illness by damaging the immune system (Marmot et al. 1997; Stewart-Brown 1998). Poor mental health in mothers is a major risk factor not only for their own Consequences of Mental Disorders and Alcohol physical ill health but also for impaired physical, cognitive, and Substance Abuse and emotional development of children and child mortality Mental disorders and alcohol and substance abuse are dis- from infectious diseases (Rutter and Quinton 1984). abling and costly. They affect productivity, impact the family and the community of the person with the disorder, and are Comorbidity. Comorbidity (the coexistence of two or associated with higher health care and other social services more disorders) has been found to be common among utilization, including the criminal justice system. The fol- patients suffering from depression. It typically involves a lowing is a brief discussion of some of the data available on combination of general physical and mental disorders or the impact of mental disorders and alcohol and substance neurological and mental disorders. In one study of patients abuse in Sub-Saharan Africa. attending primary care, of nearly 21 percent of patients with clinically significant depressive symptoms, only 1.2 percent Poverty. Psychiatric and neurological disorders impose a cited depression as the reason for their visit to the physician significant burden in developed and developing countries. (Broadhead and Abas 1994). Comorbidity of physical and In the West, considerable evidence links poverty and mental mental illness has been found to increase with age. and neurological illness. For example, the national psychi- Substance abuse is a frequent comorbid condition with atric morbidity surveys of 1993­94 in Great Britain showed depression. Studies in both developed and developing that people with any form of mental disorder had an aver- countries point to substance abuse as both a cause and effect age income of only 46.5 percent of the average income for of depression linked to both genetic and environmental the general population. In countries where there are no factors. Depression has been shown to be a major factor Mental Health and the Abuse of Alcohol and Controlled Substances | 339 Table 22.4 Selected Studies on the Psychosocial and Mental Health Consequences of HIV/AIDS Authors and year Country, study population Study findings Turner et al. 2003 United States Women less adherent than men (13% vs. 25%) 1,827 women and 3,246 men on combination drug use Women more likely to be diagnosed with depression (34% vs. 29%) Gender, depression, and ARV adherence Adherence better for those on treatment for depression de Ronchi et al. 2000 Italy 3.7% had new-onset psychosis 325 subjects, 12 with DSM-IV-R criteria for organic Generalized brain atrophy shown in CT scan of three of nine patients delusional syndrome Remission of psychotic symptoms observed in two of the new-onset psychosis patients Spire et al. 2002 France 26.7% reported nonadherence at four months of follow-up 445 patients on HAART Level of depression associated with nonadherence Other correlates: younger age, poor housing conditions, and lack of social support Morrison et al. 2002 United States, Florida 19.4% of infected women had depression compared with 4.8% of 93 HIV-infected women; 62 uninfected women seronegative women Mean scores for depression higher for the HIV-infected women HIV-positive women had higher anxiety symptom scores Evans et al. 2002 United States Major depression in 15.87% of HIV-positive women vs. 10.00% of 63 HIV-positive women; 30 HIV-negative women HIV-negative women Association of viral load in women with HIV Higher depression scores in HIV-positive women Anxiety scores similar in the two groups Depressive and anxiety scores significantly associated with higher- activated CD8 T lymphocyte cells and higher viral loads Major depression associated with lower natural killer cell activity Ciesla and Roberts 2001 2,596 with depression; 1,822 with dysthymic disorder Frequency of major depression two times higher in HIV-positive Meta-analysis of 10 studies of relation between HIV individuals infection and risk for depression Depression not related to sexual orientation or disease stage of infected individuals Drotar et al. 1999 Uganda Lower mental and motor development in HIV-infected infants 61 infants of HIV-infected mothers; 234 uninfected No group differences on mean performance or growth rates on visual infants (seroreverters) information processing 115 uninfected infants of uninfected mothers Carson et al. 1998 Kenya No substantial differences in psychiatric morbidity or neuropsychological 230 subjects functioning between the HIV-positive and HIV-negative subjects 34% HIV positive Sebit 1995 Kenya and Zaire Depression significantly higher in symptomatic seropositive individuals 408 HIV-positive individuals Determining the natural history of HIV-1 Boivin et al. 1995 Zaire Central nervous system structures affected even in seemingly asympto- 14 asymptomatic HIV-positive children under two matic HIV-positive children years of age Labor-intensive provision of care in the African milieu and global 20 HIV-negative children born to HIV-positive mothers cognitive development of even uninfected children compromised by maternal HIV-infection Bleyenheuft et al. 1992 Central African Republic HIV rate higher for those being hospitalized for the first time 292 women hospitalized for psychiatric reasons HIV appears responsible for psychiatric fragility Psychiatric symptoms apparent before the onset of symptoms of AIDS Perriens et al. 1992 Zaire Neuropsychiatric symptoms in 41% of the HIV-positive patients 196 patients studied 104 seropositive Belec et al. 1989 Central African Republic Neurologic and psychiatric symptoms in 16% 93 seropositive inpatients Source: Compiled by authors. Note: ARV antiretroviral; DSM-IV-R Diagnostic and Statistical Manual of Mental Disorders, Revised Fourth Edition; HAART highly active antiretroviral therapies. 340 | Florence K. Baingana, Atalay Alem, and Rachel Jenkins contributing to relapse in women abusing alcohol and others, and parenting. Early child interventions, as well as drugs. Identifying substance abuse in patients presenting early recognition of any problems, have been found to be with depressive illness is an important component of man- crucial to optimizing cognitive development and the future agement of the illness. Depression is also a common con- performance of children in school. comitant of HIV/AIDS (Ciesla and Roberts 2001; Morrison Strengthening communities--for example, increasing et al. 2002) and is associated with decreased compliance social inclusion and participation, improving neighborhood with medications (Turner et al. 2003), increased risk-taking environments, and developing health and social services behavior, putting others at risk for infection, and faster pro- that support mental health--and reducing discrimination gression of the course of the illness evidenced by a rapid fall and inequality by promoting access to education, meaning- in the CD4 cell counts (Baingana, Thomas, and Comblain ful occupation, and adequate housing are all appropriate 2005; Evans et al. 2002). targets for promoting mental health. Addressing stigma in Patients with epilepsy often present with psychological relation to mental disorders is also crucial. or psychotic symptoms. Sixty percent of 230 patients with epilepsy who were referred to the neurology clinic of Primary Prevention Muhimbili Medical Center, Tanzania, had a psychological Prevention is critical in reducing the impact of mental disturbance warranting intervention; 81 percent had a disorders and alcohol and substance abuse. On ethical minor neurotic disorder, but 19 percent had schizophreni- grounds alone, prevention is always preferable to treatment form psychosis (Matuja 1990). Other disturbances were or rehabilitation. In most instances prevention is also more agoraphobia and severe depression. Over 80 percent of cost-effective than treatment. Many potentially catastroph- patients with a major disturbance had epilepsy and a brain ic disorders are now preventable. Examples include iodine lesion; 77 percent of patients with a minor disturbance had supplementation to prevent mental retardation and iodine evidence of an organic brain lesion. Organic brain lesion deficiency disorder, as well as immunization against and psychological disturbance were overwhelmingly associ- tetanus, tuberculosis, measles, rubella, and polio; immu- ated with social disadvantage. nization in the perinatal and early child period prevents In a study of 478 Zambian patients on a given day, all of these infections, which can damage the central nervous whom were examined for goiter, 34.4 percent of all adult system and could have epilepsy as a sequelae, and mental females and 23.2 percent of all adult males were found to retardation (Down's Syndrome) is associated with rubella have goiter (Rwegellera and Mambwe 1977). Goiter was in the first trimester of pregnancy. Zinc, folic acid, and iron found in 57.6 percent of females with affective illness and supplementation and fortification are crucial to preventing 77 percent of males with paranoid psychosis. Bademosi and learning disabilities, mental retardation, and developmen- colleagues (1976) found that 38 percent of patients with tal delays. Prevention of maternal transmission of HIV infective endocarditis had neuropsychiatric symptoms. For is increasingly critical, thus preventing the neurological 75 percent of these, the neuropsychiatric symptoms were the and developmental consequences associated with pediatric presenting feature. HIV infection. Safe motherhood initiatives, such as atten- dance at an antenatal clinic, tetanus vaccination during pregnancy, and delivery with a trained attendant, also greatly INTERVENTIONS reduce the impact of prenatal, perinatal, and postnatal risk Possible interventions available and feasible in the context of factors. Sub-Saharan Africa are grouped here into the traditional Preventive community-wide psychosocial programs three: promotion, prevention, and treatment. have been shown to be effective, especially for populations affected by conflict and HIV/AIDS. These include child care centers for orphans and vulnerable children, children's Promotion clubs, school-based mental health programs, and support The West has provided evidence about the value of inter- groups in the communities. ventions to strengthen individuals' mental well-being and Recently, short-course antiretroviral prophylaxis regi- increase emotional resilience. Such interventions are mens have been shown to provide a relatively low cost and designed to promote self-esteem and improve life and cop- effective strategy for preventing mother-to-child transmis- ing skills, including communicating, negotiating, relating to sion of HIV in low-income populations by up to 30 percent Mental Health and the Abuse of Alcohol and Controlled Substances | 341 (Connor et al. 1994; Mofenson 1999). Supplemental feeding primary care level across Sub-Saharan Africa still represents reduces the transmission through breastfeeding by a further a major financial challenge for these countries, even though 30 percent, thus preventing the neurological and develop- TCAs are relatively cheaper than the SSRIs, even if the latter mental disabilities associated with HIV infection in children. are now getting off patent protection. The median yearly cost for treating depression in an individual with amitriptyline in Sub-Saharan Africa was estimated at US$30.66 in 2001 and Treatment is now estimated to be US$34.38 (WHO 2001, 2005). A brief description is provided here for the treatments avail- Bolton and colleagues (2003) tested the efficacy of group able for the mental disorders discussed. A concerted effort interpersonal psychotherapy in alleviating depression and was made to provide treatment alternatives where the costs dysfunction in rural Uganda and found it to be highly and effectiveness are known, but this is not always possible efficacious. Mean reduction of depression severity was for Sub-Saharan Africa, since these studies are just being 17.47 points for intervention groups and 3.55 points for con- carried out. The next best alternative is to provide a discus- trols. After the intervention, 6.5 percent of the intervention sion of what is presently being provided as treatment, even group and 54.7 percent of the control group met the criteria when evidence for cost-effectiveness may not be available. for major depression compared with 86 percent and 94 per- cent, respectively, before treatment was initiated. Depression. Effective treatment strategies exist for depres- Cognitive behavior therapy, problem-solving therapy, sion in the form of pharmacological agents, cognitive and family-focused therapy have met with success in the behavioral therapy, and psychosocial treatments. Although treatment of depression. A small number of published treatment interventions may not cure all forms of depres- reports address the use of psychosocial interventions to treat sion, a large number of efficacious and low-cost treatments depression in developing countries. Problem-solving therapy are available. Despite the availability of these interventions, has been suggested as an effective psychosocial treatment, many people in Africa remain undiagnosed and untreated. particularly because it seeks to provide the patient with a It is difficult to estimate the actual treatment gap (all those technique for coping with future problems, thereby poten- with a clinically diagnosed mental disorder who are not on tially preventing a recurrence of depressive symptoms or treatment) because epidemiological data of diagnosed men- enabling the patient to deal with them more effectively when tal disorders in the community or of those who are on they recur. Problem-solving therapy has been conducted treatment are limited. The scant data available are often effectively by trained community nurses in primary care unreliable. Andrews, Henderson, and Wayne-Hall (2001), settings, making the approach particularly attractive for studying utilization of the Australian mental health services, resource-poor settings, where psychiatrists and specially found that only about 30 percent of those with a diagnosed trained general physicians are not available. mental disorder used the services. Because of their efficacy and cost-effectiveness, antide- Schizophrenia. Evidence suggests that correct early diag- pressant medications represent the mainstay of treatment nosis and initiation of treatment can have a positive impact for depression in developed countries. Seventy percent of on the subsequent course of schizophrenia. Antipsychotic patients prescribed antidepressants show significant clinical medication is the mainstay of treatment and is indicated for improvement. Antidepressants are also effective as prophy- the majority of patients over prolonged periods with no laxes: treatment has been shown to reduce the relapse rate fixed limit to duration. Two classes of pharmacological for recurrent depression from 80 percent over three years to agents are available. The two offer approximately equal effi- 22 percent. There has been far less research in developing cacy in controlling the positive symptoms of the disorder countries, but the limited available evidence shows similar but differ considerably in their side effects and tolerability rates of efficacy. Tricyclic antidepressants (TCAs) and the as well as cost. The median yearly cost per person for chlor- newer selective serotonin reuptake inhibitors (SSRIs) promazine for Sub-Saharan Africa was estimated at have similar efficacy for moderate depression. The reduced US$40.88 in 2001 and is now estimated to be US$49.06 side effects of SSRIs enhance patient compliance. However, (WHO 2001, 2005). the high cost of SSRIs means that they remain out of reach as a first-line treatment in Africa for all but the wealthy. Posttraumatic Stress Disorder. A combination of psy- Indeed, simply ensuring an adequate supply of TCAs to the chosocial and mental health interventions is recommended 342 | Florence K. Baingana, Atalay Alem, and Rachel Jenkins for PTSD. Psychosocial interventions include counseling, Butajira community laboratory (Alem et al. 1999; Kebede group support meetings, play activities, art, music, and et al. 2003; Kebede et al. 2006), where a group of Ethiopian other expressive art therapies. Mental health interventions researchers is conducting cross-sectional and longitudinal include a short course of anxiolytics for acute distress, not to mental health studies could be cited as a good example for be taken for longer than two weeks. Symptoms that persist such undertakings in Sub-Saharan Africa. beyond the acute phase respond to smaller doses of antide- · Multisite studies on the mental well-being of children pressants and antipsychotics. Drug treatment for PTSD is would establish the incidence and prevalence of mental best combined with a psychotherapeutic intervention, such disorders of children, links to abuse of all forms, as group therapy, individual therapy, or counseling. malnutrition, conflicts, poverty and vulnerability, what interventions are available, what the burden is within the Mental Disorders among Children. The first step in the school system, and how it affects educational outcomes. management of mental disorders among children is making · Stigmatization confounds epidemiological studies, pre- the correct diagnosis. Management is also dependent on a vents treatment, and leads to personal and economic dis- collaboration between the parents or caretaker, the teachers, aster for many affected individuals and their families. and the health care provider. Treatment depends on the Because stigma and other cultural beliefs are locally diagnosis and the underlying causative factors. Antidepres- grounded, they require local study in order to develop sants are effective in the management of emotional disor- ways in which to deal with them constructively. ders of children. However, the teacher must be aware of the · Cost-of-illness, cost-effectiveness, cost-minimization, diagnosis and provide support within the school system. For and cost-benefit analyses should be undertaken to pro- children with learning difficulties, special needs education vide government officials and funding agencies the teachers play a crucial role in providing education tailored necessary economic perspective in relation to mental dis- to the needs of the child. orders and alcohol abuse. Socioeconomic determinants For children with attention­deficit/hyperactivity disor- of the outcome measures that could best reflect optimal der, Ritalin and other stimulants are not widely used in Sub- care for people with mental disorders and those abusing Saharan Africa, mainly because of the lack of child psychia- alcohol should be identified. trists and psychologists, necessary for the close supervision · Family, twin, and adoption studies in Western countries required. High activity levels can be managed with behav- have provided evidence of the genetic contribution to the ioral methods, and a special needs education teacher can etiology of depressive disorders. Twin studies have also design a learning program tailored to the attention deficit. demonstrated the strong role of the nonshared environ- A few countries, such as Kenya, South Africa, and Uganda, ment in the causation of depression. Thus, examination have developed comprehensive special education teacher- of gene-environment interactions is essential to future training programs. research and would benefit from a wide range of varia- tion in the psychosocial environments, making it possible A FRAMEWORK FOR RESEARCH to delineate better the etiological contributions of such environmental factors as socioeconomic adversity, life In developing short-term, medium-term, and long-term events, and the breakdown of social support networks. strategies, it is clear that further research will have to be There is a need for such studies to be carried out in Sub- carried out in order to provide the evidence necessary to Saharan Africa. strengthen the mental health care systems of Sub-Saharan · Defective neurotransmission and neuroendocrine recep- Africa. Following are some of the recommended research tor responses are associated with depression. Three mo- areas that could be pursued. noamines have been implicated: 5-hydroxytryptamine (serotonin), noradrenaline, and dopamine. Conclusive · Cross-sectional and longitudinal studies of mental disor- evidence on the impact of these defects on the course of ders in Sub-Saharan Africa, including validation of the depressive disorders remains to be found. It is still standardized testing instruments for the different popula- unclear whether associations with neurological function tions of Africa, would establish the epidemiology and represent cause or effect in the pathogenesis of depres- causative and risk factors, as well as links to sexual abuse, sion and whether neurological function is the major risk violence against women, HIV/AIDS, and conflicts. The factor for depression in Sub-Saharan Africa. Mental Health and the Abuse of Alcohol and Controlled Substances | 343 · Additional research should be conducted in developing excess prevalence rates of preventable mental disorders-- countries to determine the cost-effectiveness of cognitive such as depression, PTSD, and mental disorders resulting behavioral therapy in primary health care settings. from sexual abuse of children--observed in Sub-Saharan Requisite long-term follow-up studies have not been Africa countries today is a consequence of both poverty conducted in low-income countries. They would need to and poor resource allocation, and it is an impediment to include attention to recovery from symptoms, dysfunc- future social and economic development. tion and reversal of the dysfunction following treatment, · To implement effective programs for community-level disability, the family's support of the patient, and the cost detection and treatment of patients with mental disor- of this support to the family, as well as the social and ders, governments and health authorities of Sub-Saharan economic burden on the family and the community. Africa countries must first issue clear policies articulat- Measurements would have to be made over at least five ing measures for the identification and treatment of years. patients with mental disorders, including children. · Not enough evaluations of mental health services have Essential components of any strategy include the assur- been carried out in Sub-Saharan Africa, and the few that ance of a continuous drug supply and means of distribu- have been done are neither comprehensive nor country- tion, prevention programs, social programs, and stigma wide. The process for integrating mental health into pri- mitigation strategies to increase employment and mary health care and the effectiveness of this approach improve general welfare among individuals suffering needs to be documented. from the disease. Also essential are training of health care personnel and methods for surveillance and routine data collection. Reforms in the health, social, and economic POLICY ISSUES sectors need to address mental health and substance abuse clearly and in an integrated way, and programs The policy implications resulting from the epidemiology need to be set up targeting other priority areas, such as of mental health and substance abuse disorders in Sub- HIV, malaria and other infectious diseases, and repro- Saharan Africa discussed here are provided only as a menu ductive and child health. of possible policy interventions, to be selected from, depending on the context and the needs of each country. Increasing the policy and service development and the clinical and research professional capacity in Sub-Saharan · As societies and economies become increasingly infor- African countries and stemming the flow of skilled health mation oriented and dependent on highly skilled and professionals to wealthy countries (Hongoro and McPake literate workers, it is critical that children everywhere 2004; Liese, Blanchet, and Dussault 2003; WHO, NEPAD, have an opportunity to reach their optimal levels of cog- and ACOSHED 2005) are key to developing sustainable, nitive and neurological development. The persistence of locally appropriate programs. 344 | Florence K. Baingana, Atalay Alem, and Rachel Jenkins APPENDIX 22A Table 22A.1 Selected Sub-Saharan Africa Data on Mental Health Disorders Author(s) Disorder Country, study population Findings Molteno et al. 2001 Behavioral and Cape Town, South Africa 31% for psychopathology, boys more affected emotional disorders 355 children in special schools than girls and intellectual More behavioral problems among the children with disability severe and profound retardation Epilepsy associated with more total behavior scores Ikeji et al. 1999 ECT and schizophrenia, Nigeria Full clinical recovery mania, and severe 70 subjects Unmodified ECT safe and effective with nonenduring depression Prospective open-label study subjective memory difficulty Bolton, Neugebauer, and Depression Rwanda 15.5% met criteria for DSM IV diagnosis of Ndogoni 2002 368 adults depression. Community-based random sample Kaaya et al. 2002 Depression among Tanzania Internal consistency of HSCL-25 adequate HIV-positive women 903 women HSCL-25 demonstrated utility for depression Two-phase design Bean and Moller PTSD and depression South Africa 63% moderately to severely depressed 2002 40 battered women 59% high PTSD symptoms 38.4% anger, 54.5% guilt Martenyi et al. 2002 PTSD Europe, Israel, and South Africa Fluoxetine associated with greater improvement Double-blind, randomized, Fluoxetine effective and well tolerated in PTSD placebo-controlled study Efficacy and tolerability of fluoxetine 226 patients on fluoxetine, 75 on placebo Bolton 2001b Mental health effects Rwanda Depression occurs in this population of genocide Free listing, key informant Supports local content validity of the depression interviews, and pile sorts assessment instruments Njenga 2000 Depression Kenya 22% reported depressive symptoms 86 professional women 30% coping less well than usual Mkize, Nonkelela, and Depression Transkei, South Africa 53% mild to severe depression Mkize 1998 250 students randomly selected Females more affected 3 to 1 Beck's Depression Inventory 14% moderately to severely depressed All subjects presented with somatic symptoms Vaz, Mbajiorgu, and Stress, depression, and suicide Zimbabwe 64.5% at various levels of stress or depression Acuda 1998 109 medical students or both Cross-sectional study 11% very high levels of stress 12% at serious risk for suicide Lopes and Bottino 2002 Dementia All continents 1.17% specific prevalence rate for dementia Medline and Lilacs search, for 65­69-year-olds 38 studies evaluated from all 54.83% specific prevalence rate for dementia for continents those over 95 years Dementia more prevalent among women in 75% papers reviewed Aina 2001 Clinical profile of patients Nigeria Highest percentage made up of young adults attending psychiatric Prospective, private hospital­based 31­45 years of age hospital study 36% epilepsy 138 patients seen in 22.5% schizophrenia 644 consultation sessions 18.8% affective disorders (Table continues on the next page.) Mental Health and the Abuse of Alcohol and Controlled Substances | 345 Table 22A.1 (Continued) Author(s) Disorder Country, study population Findings Molteno et al. 2001 Behavioral and emotional South Africa 31% psychopathology in children with intellectual problems in children with 355 children with intellectual disability intellectual disability disability attending special More behavioral problems in boys than girls schools More behavioral difficulties in children with severe and profound forms of intellectual disability Kwalombota 2002 HIV and depression Zambia 85% of HIV-positive women had major depressive Mental health of HIV-positive disorder women attending antenatal clinic Depression more common among those diagnosed HIV positive during pregnancy Nwosu and Odesanmi 2001 Suicide Nigeria Suicides at the rate of 0.4 per 100,000 population Study of pattern of autopsy findings Higher suicide incidence in males (3.6 to 1) in cases of completed suicides Majority of the victims in their 20s Bhagwanjee et al. 1998 Minor psychiatric South Africa 23.9% prevalence of generalized anxiety and disorders 354 adults depressive disorder Two-stage community-based 3.7% generalized anxiety, 4.8% major depression, epidemiological study 7.3% dysthymia, and 8.2% major depression with dysthymia Okulate 2001 Suicide Nigeria Suicide 0.37% of all admissions to the Department Case-control study of the character- of Psychiatry, Military Hospital, Yaba, Nigeria istics of patients who attempted 60.8% of all suicides below the age of 30 years suicide in a military setting Numbers of males and females almost equal 51 attempted suicides Dong and Simon 2001 Organophosphate poisoning Zimbabwe Increase of 320% in organophosphate poisoning Cross-sectional descriptive study between 1995 and 2000 of the use of organophosphate Similar male and female admission rates as poison 82% below 31 years Urban hospital admissions 74% suicide attempts 183,569 case records studied 599 cases of organophosphate poisoning Kebede and Alem 1999 Suicide Ethiopia Prevalence of current suicide ideation, 2.7% Study of suicide attempts and Lifetime prevalence, 0.9% suicide ideation 66% of subjects below 25 years of age 10,203 adults in Addis Ababa Current suicide ideation more common in men than women (95% confidence interval) Hanging preferred method for men, poisoning for women Ihueze and Okpara Psychiatric disorders of Nigeria Patients over 60 years admitted for the first time, 1989 old age Retrospective study of 73 consecu- 5% of all admissions tive patients age 60 years and 58% below 70 years over admitted to a psychiatric 84% in the two lowest socioeconomic classes hospital 49% functional psychosis 30% organic psychosis 10% neurotic disorders Ben-Arie et al. 1983 Psychiatric disorders of South Africa 24% some form of psychiatric disorder old age 139 noninstitutionalized coloured 16.5% depression persons over 65 years old 15% alcoholism among the men 6% on psychiatric medication Verrier-Jones et al. 1978 Psychiatric disorders of South Africa Over 50% of patients depressed, many associated old age 100 patients admitted to a with physical illness and isolation psychogeriatric unit Nine patients admitted with confusion Confusion due to drugs prescribed by medical practitioners in seven of them Source: Compiled by authors. Note: ECT electroconvulsive therapy. 346 | Florence K. Baingana, Atalay Alem, and Rachel Jenkins NOTES Ben-Arie, O., L. Swartz, A. F. Teggin, and R. Elk. 1983. "The Coloured Elderly in Cape Town--A Psychosocial, Psychiatric and Medical 1. A demographic study site is a geographical area that is delineated Community Survey." Part 2. Prevalence of Psychiatric Disorders." and, after the baseline assessment of the total population is entered into a South African Medical Journal 64 (27): 1056­61. database, used for regular surveillance of health and demographic trends Berard, R. M. F., and F. Boermeester. 1999. "Sexual Abuse in Adolescents: over time. Data from a Psychiatric Treatment Center for Adolescents." South 2. AFR D: Algeria, Angola, Benin, Burkina Faso, Cameroon, Cape African Medical Journal 89 (9): 972­76. Verde, Chad, Comoros, Equatorial Guinea, Gabon, Gambia, Ghana, Bhagwanjee, A., A. Parekh, Z. Paruk, I. Petersen, and H. Subedar. 1998. Guinea, Guinea-Bissau, Liberia, Madagascar, Mali, Mauritania, Mauritius, "Prevalence of Minor Psychiatric Disorders in an African Rural Niger, Nigeria, São Tomé and Principe, Senegal, Seychelles, Sierra Leone, Community in South Africa." Psychological Medicine 28: 1137­47. Togo. AFR E: Botswana, Burundi, Central African Republic, Côte d'Ivoire, Bleyenheuft, L., P. Janne, C. Reynaert, N. Munyandamutsa, F. Verhoeven, B. Democratic Republic of Congo, Eritrea, Ethiopia, Kenya, Lesotho, Malawi, Muremyangango, and L. Cassiers. 1992. "Prevalence of Human Mozambique, Namibia, Republic of Congo, Rwanda, South Africa, Immunodeficiency Virus Infection in a Psychiatric Population in Swaziland, Uganda, Tanzania, Zambia, Zimbabwe. Central Africa." Acta Psychiatrica Belgica 92 (2): 99­108. Boivin, M. J., S. D. Green, A. G. Davies, B. Giordani, J. K. Mokili, and W. A. Cutting. 1995. "A Preliminary Evaluation of the Cognitive and Motor Effects of Pediatric HIV Infection in Zairian Children." Health REFERENCES Psychology 14 (1): 13­21. Bolton, P. 2001a. "Cross-Cultural Validity and Reliability Testing of a Abas, M., and J. Broadhead. 1994. "Mental Disorders in the Developing Standard Psychiatric Assessment Instrument without a Gold World." British Medical Journal 308: 1052­53. Standard." Journal of Nervous and Mental Disease 189 (4): 238­42. ------. 1997. "Depression and Anxiety amongst Women in an Urban ------. 2001b. "Local Perceptions of the Mental Health Effects of the Setting in Zimbabwe." Psychology and Medicine 27: 59­71. Rwandan Genocide." Journal of Nervous and Mental Disease 189 (4): Abdulahi, H., D. H. Mariam, and D. Kebede. 2001. "Burden of Disease 243­48. Analysis in Rural Ethiopia." Ethiopian Medical Journal 39: 271­81. Bolton, P., J. Bass, R. Neugebauer, H. Verdeli, K. Clougherty, P. Aina, F. 2001. "Socio-Demographic and Clinical Profile of Patients Wickramatne, L. Speelman, L. Ndogoni, and M. Weissman. 2003. Attending a Private Psychiatric Hospital in Lagos, Nigeria." West "Group Interpersonal Psychotherapy for Depression in Rural Uganda." African Journal of Medicine 20 (2): 117­22. Journal of the American Medical Association 289 (23): 3117­24. Alem, A., D. Kebede, L. Jacobsson, and G. Kullgren. 1999. "Suicide Bolton, P., R. Neugebauer, and L. Ndogoni. 2002. "Prevalence of Attempts among Adults in Butajira, Ethiopia." Acta Psychiatrica Depression in Rural Rwanda Based on Symptom and Functional Scandinavica 397 (Suppl.): 70­76. Criteria." Journal of Nervous and Mental Disease 190 (9): 631­37. Andrews, G., S. Henderson, and W. Wayne-Hall. 2001. "Prevalence, Booyens, J., M. L. Luitingh, and C. F. van Rensburg. 1977. "The Relation- Comorbidity, Disability and Service Utilization: Overview of the ship between Scholastic Progress and Nutritional Status. Part 2. A One Australian National Mental Health Survey." British Journal of Psychiatry Year Follow-Up Study." South African Medical Journal 52 (16): 650­52. 178: 145­53. Bouta, T., G. Frerks, and I. Bannon. 2004. Gender, Conflict and Bademosi, O., A. O. Falasae, F. Jaiyesimi, and A. Bademosi. 1976. Development. Washington, DC: World Bank. "Neuropsychiatric Manifestations of Infective Endocarditis: A Study of Breetzke, K. A. 1988. "Suicide in Cape Town--Is the Challenge Being Met 95 Patients at Ibadan, Nigeria." Journal of Neurology, Neurosurgery and Effectively?" South African Medical Journal 73 (1): 19­52. Psychiatry 39 (4): 325­29. Broadhead, J., and M. Abas. 1994. "Depressive Illness--Zimbabwe." Baingana, F., I. Bannon, and R. Thomas. 2005. "Mental Health and Tropical Doctor 24 (1): 27­30. Conflicts: Conceptual Framework and Approaches." A Health, Carson, A. J., R. Sandler, F. N. Owino, F. O. Matete, and E. C. Johnstone. Nutrition and Population Discussion Paper, World Bank, Washington, 1998. "Psychological Morbidity and HIV in Kenya." Acta Psychiatrica DC. Scandinavica 97 (4): 267­71. Baingana, F., A. Dabalen, E. Menye, M. Prywes, and M. Rosholm. 2004. CDC (Centers for Disease Control and Prevention). 2000. "Cause-Specific "Mental Health and Socio-Economic Outcomes in Burundi." A Health, Adult Mortality: Evidence from Community-Based Surveillance-- Nutrition and Population Discussion Paper, World Bank, Washington, Selected Sites, Tanzania, 1992­1998." Morbidity and Mortality Weekly DC. Report 19: 49. Baingana, F., R. Thomas, and C. Comblain. 2005. "HIV and Mental Ciesla, J. A., and J. E. Roberts. 2001. "A Meta-Analysis of Risk for Major Health." A Health, Nutrition and Population Discussion Paper, World Depressive Disorder among HIV-Positive Individuals." American Bank, Washington, DC. Journal of Psychiatry 15: 725­30. Barton, T., and A. Mutiti. 1998. Northern Uganda Psycho-Social Needs Connor, E. M., R. S. Sperling, R. Gelber, P. Kiselev, G. Scott, M. J. O. Assessment (NUPSNA). Geneva: UNICEF. O'Sullivan, et al. 1994. "Reduction of Maternal-Infant Transmission of Bean, J., and A. T. Moller. 2002. "Posttraumatic Stress and Depressive Human Immuno-Deficiency Virus Type 1 with Zidovudine Symptomatology in a Sample of Battered Women from South Africa." Treatment." New England Journal of Medicine 331 (180): 1173­80. Part 1. Psychological Reports 90 (3): 750­52. Cummins, R. R., and C. W. Allwood. 1984. "Suicide Attempts or Threats by Beck, A. T. 1986. "Hopelessness as a Predictor of Eventual Suicide." Annals Children and Adolescents in Johannesburg." South African Medical of the New York Academy of Sciences 48: 90­96. Journal 66 (19): 726­29. Belec, P. L., J. Testa, M. D. Vohito, G. Gresenguet, M. I. Martin, A. Tabo, de Jong, J. 2002. "Public Mental Health, Traumatic Stress and Human B. Di Costanzo, and A. J. Georges. 1989. "Neurologic and Psychiatric Rights Violations in Low-Income Countries." In Trauma, War and Manifestations of AIDS in Central African Republic." Bulletin de la Violence: Public Mental Health in Socio-Cultural Context. Plenum Societé de Pathologie Exotique et de Ses Filiales 82 (3): 297­307. Series in Social/Clinical Psychology. New York: Plenum. Mental Health and the Abuse of Alcohol and Controlled Substances | 347 Demyttenaere, K., R. Bruffaerts, J. Posada-Villa, I. Gasquet, V. Kovess, J. P. Jablensky, A., N. Sartorius, G. Ernberg, M. Anker, A. Korten, J. E. Cooper, Lepine, M. C. Angermeyer, et al. 2004. "Prevalence, Severity, and R. Day, and A. Bertelsen. 1992. "Schizophrenia: Manifestations, Unmet Need for Treatment of Mental Disorders in the World Health Incidence and Course in Different Cultures. A World Health Organization World Mental Health Surveys." Journal of the American Organization Ten-Country Study." Psychology and Medicine 20: Medical Association 291 (21): 2581­90. 1­97. de Ronchi, D., I. Faranca, P. Forti, G. Ravaglia, M. Borderi, R. Manfredi, and V. Jenkins, R. 1985. "Minor Psychiatric Morbidity in Employed Young Men Volterra. 2000. "Development of Acute Psychotic Disorders and HIV-1 and Women and Its Contribution to Sickness Absence." British Journal Infection." International Journal of Psychiatry in Medicine 30 (2): 173­83. of Industrial Medicine 42 (3): 147­54. de Villiers, F. P. R., and M. A. Prentice. 1996. "Accumulating Experience in ------. 1998. "Linking Epidemiology and Disability Measurement with a Child Abuse Clinic." South African Medical Journal 86 (2): 147­50. Mental Health Service Policy and Planning." Epidemiologia e Dhadphale, M., and B. Ibrahim. 1984."Learning Disabilities among Nairobi Psichiatria Sociale 7 (2): 120­26. School Children." Acta Psychiatrica Scandinavica 69 (2): 151­55. Kaaya, S. F., M. C. Fawzi, J. K. Mbwambo, B. Lee, G. I. Msamanga, and W. Fawzi. 2002. "Validity of the Hopkins Symptom Checklist-25 Dong, X., and M. A. Simon. 2001."The Epidemiology of Organophosphate amongst HIV-Positive Pregnant Women in Tanzania." Acta Psychiatrica Poisoning in Urban Zimbabwe from 1995 to 2000." International Scandinavica 106 (1): 9­19. Journal of Occupational and Environmental Health 7 (4): 333­38. Kebede, D., and A. Alem. 1999. "Suicide Attempts and Ideation among Drotar, D., K. Olness, M. Wiznitzer, C. Schatschneider, L. Marum, L. Guay, Adults in Addis Ababa, Ethiopia." Acta Psychiatrica Scandinavica 397: J. Fagan, et al. 1999. "Neurodevelopmental Outcomes of Ugandan 35­39. Infants with HIV Infection: An Application of Growth Curve Analysis." Health and Psychology 18 (2): 114­21. Kebede, D., A. Alem, T. Shibire, N. Deyassa, A. Negash, T. Beyero, G. Medhin, and A. Fekadu. 2006. "Symptomatic and Functional Outcome Eferakeya, A. E. 1984. "Drugs and Suicide Attempts in Benin City, Nigeria." of Bipolar Disorder in Butajira, Ethiopia." Journal of Affective Disorders British Journal of Psychiatry 145: 70­73. 90 (2­3): 239­49. Evans, D. L., T. R. Ten Have, S. D. Douglas, D. R. Gettes, M. Morrison, Kebede, D., A. Alem, T. Shibire, A. Negash, A. Fekadu, D. Fekadu, L. M. S. Chiappini, et al. 2002. "Association of Depression with Viral Jacobsson, and G. Kullgren. 2003. "Onset and Clinical Course of Load, CD8 T Lymphocytes, and Natural Killer Cells in Women with Schizophrenia in Butajira-Ethiopia: A Community Based Study." Social HIV Infection." American Journal of Psychiatry 159: 1752­59. Psychiatry and Psychiatric Epidemiology 38: 625­31. Flisher, A. J., and C. D. Parry. 1994. "Suicide in South Africa. An Analysis of Kebede, D., and T. Ketsela. 1993. "Suicide Attempts in Ethiopia in Addis Nationally Registered Mortality Data for 1984­1986." Acta Psychiatrica Abeba High School Adolescents." Ethiopian Medical Journal 31: 83­99. Scandinavica 90 (5): 348­53. Kessler, R. C., W. T. Chiu, O. Demler, K. R. Merikangas, and E. E. Walters. Friedman, E. 2004. "Mental Health Effects of the Indonesian Economic 2005. "Prevalence, Severity, and Comorbidity of 12-Month DSM-IV Crisis." Development Economics Group Discussion Paper, World Disorders in the National Comorbidity Survey Replication." Archives of Bank, Washington, DC. General Psychiatry 62 (6): 617­27. Goldberg, D., and P. Huxley. 1992. Common Mental Disorders. London: Kessler, R. C., K. A. McGonagle, S. Zhao, C. B. Nelson, M. Hughes, S. Routledge. Eshleman, H. U. Wittchen, and K. S. Kendler. 1994. "Lifetime and Goldberg, D. P., R. Gater, N. Sartorius, T. B. Ustun, M. Piccinelli, O. Gureje, 12-Month Prevalence of DSM-III-R Psychiatric Disorders in the and C. Rutter. 1997. "The Validity of Two Versions of the GHQ in the United States. Results from the National Comorbidity Survey." Archives WHO Study of Mental Illness in General Health Care." Psychology and of General Psychiatry 51 (1): 8­19. Medicine 27: 191­97. Kiima, D. M., F. G. Njenga, M. M. Okonji, and P. A. Kigamwa. 2004. "Kenya Granja, A. C., E. Zacarias, and S. Bergstrom. 2002. "Violent Deaths: The Mental Health Country Profile." International Review of Psychiatry Hidden Face of Maternal Mortality." British Journal of Obstetrics and 16 (1­2): 48­53. Gynaecology 109 (1): 5­8. Kwalombota, M. 2002. "The Effect of Pregnancy in HIV-Infected Women." Green, B. L., M. J. Friedman, J. de Jong, S. D. Solomon, T. M. Keane, J. A. AIDS Care 14 (3): 431­33. Fairbank, B. Donelan, and E. Frey-Wouters. 2003. Trauma Leff, J., N. Sartorius, A. Jablensky, A. Korten, and G. Ernberg. 1992. "The Interventions in War and Peace: Prevention, Practice, and Policy. New International Pilot Study of Schizophrenia: Five-Year Follow-Up York: Kluwer Academic/Plenum. Findings." Psychological Medicine 22 (1): 131­45. Hippisley-Cox, J., K. Fielding, and M. Pringle. 1998. "Depression as a Risk Lett, H. S., J. A. Blumental, M. A. Babyak, A. Sherwood, T. Strauman, Factor for Ischaemic Heart Disease in Men: Population Based Case- C. Robins, and M. F. Newman. 2004. "Depression as a Risk Factor for Control Study." British Medical Journal 316: 1714­19. Coronary Artery Disease: Evidence, Mechanisms, and Treatment." Hollifield, M., W. Katon, D. Spain, and L. Pule. 1990. "Anxiety and Psychosomatic Medicine 66 (3): 305­15. Depressions in a Village in Lesotho, Africa: A Comparison with the Liese, B., N. Blanchet, and G. Dussault. 2003. "Background Paper on the United States." British Journal of Psychiatry 156: 343­50. Human Resource Crisis in Health Services," World Bank, Washington, Hongoro, C., and B. McPake. 2004. "How to Bridge the Gap in Human DC. Resources for Health." Lancet 364: 1451­56. Lopes, M. A., and C. M. Bottino. 2002. "Prevalence of Dementia in Several Ihezue, U. H., and E. Okpara. 1989. "Psychiatric Disorders of Old Age in Regions of the World: Analysis of Epidemiologic Studies from 1994 to Enugu, Nigeria. Sociodemographic and Clinical Characteristics." Acta 2000." Arquivos de Neuro-Psiquiatria 60 (1): 61­69. Psychiatrica Scandinavica 79 (4): 332­37. Maj, M., R. Janssen, F. Starace, M. Zaudig, P. Satz, B. Sughondhabirom, Ikeji, O. C., J. U. Ohaeri, R. O. Osahon, and R. O. Agidee. 1999."Naturalistic M. A. Luabeya, et al. 1994. "WHO Neuropsychiatric AIDS Study, Comparative Study of Outcome and Cognitive Effects of Unmodified Cross-Sectional Phase 1. Study Design and Psychiatric Findings." Electro-Convulsive Therapy in Schizophrenia, Mania and Severe Archives of General Psychiatry 51 (1): 39­49. Depression in Nigeria." East African Medical Journal 76 (11): 644­50. Marais, A., P. J. T. De Villiers, A. T. Moller, and D. Stein. 1999. "Domestic IOM (Institute of Medicine). 2001. Neurological, Psychiatric, and Violence in Patients Visiting General Practitioners, Prevalence, Developmental Disorders: Meeting the Challenges in the Developing Phenomenology and Association with Psychopathology." South African World. Report. Washington, DC: IOM. Medical Journal 89 (6): 635­40. 348 | Florence K. Baingana, Atalay Alem, and Rachel Jenkins Marmot, M., H. Bosma, H. Hemingway, E. J. Brunner, and S. A. Stansfield. Njenga, F. 2000. "Depression in Kenyan Professional Women." Inter- 1997. "Contribution of Job Control and Other Risk Factors to Social national Clinical Psychopharmacology 15 (Suppl. 3): S35­36. Variations in Coronary Heart Disease Incidence." Lancet 350: 235­39. Nwosu, S. O., and W. O. Odesanmi. 2001. "Pattern of Suicides in Ile-Ife, Martenyi, F., E. B. Brown, H. Zhang, S. C. Koke, and A. Prakash. 2002. Nigeria." West African Journal of Medicine 20 (3): 259­62. "Fluoxetine v. Placebo in Prevention of Relapse in Post-Traumatic Odejide, A. O., A. O. Williams, J. U. Ohaeri, and B. A. Ikuesan. 1986. "The Stress Disorder." British Journal of Psychiatry 181: 315­20. Epidemiology of Deliberate Self-Harm: The Ibadan Experience." Mathers, C. D., C. Stein, D. Ma Fat, C. Rao, M. Inoue, N. Tomijima, C. British Journal of Psychiatry 149: 734­37. Bernard, A. D. Lopez, and C. J. L. Murray. 2002. "Global Burden of Oguleye, A. O., G. B. Nwaorgu, and H. Grandawa. 2002. "Corrosive Disease 2000: Version 2, Methods and Results." Global Program on Oesophagitis in Nigeria: Clinical Spectrums and Implications." Evidence for Health Policy Discussion Paper 50, WHO, Geneva. Tropical Doctor 32 (2): 78­80. Matuja, W. B. 1990. "Psychological Disturbance in African Tanzanian Okulate, G. T. 2001. "Suicide Attempts in a Nigerian Military Setting." East Epileptics." Tropical and Geographical Medicine 42 (4): 359­64. African Medical Journal 78 (9): 493­96. Mboussou, M., and L. Milebou-Aubusson. 1989. "Suicides and Attempted Orley, I., D. M. Blitt, and I. K. Wing. 1979. "Psychiatric Disorders in Two Suicides at the Jeanne Ebori Foundation, Libreville (Gabon)." Medicina African Villages." Archives of General Psychiatry 36: 513­21. Tropical 49 (3): 259­64. Ormel, J., M. Von Korff, T. B. Ustun, S. Pini, A. Korten, and T. Oldehinkel. Mkize, L. P., N. F. Nonkelela, and D. L. Mkize. 1998. "Prevalence of 1994. "Common Mental Disorders and Disability across Cultures. Depression in a University Population." Curationis 21 (3): 32­37. Results from the WHO Collaborative Study on Psychological Problems Mofenson, L. M. 1999. "Short Course Zidovudine for Prevention of in General Health Care." Journal of the American Medical Association Perinatal Infection." Lancet 353: 766­67. 272 (22): 1741­48. Mollica, R., K. McInnes, C. Poole, and S. Tor. 1998. "Dose-Effect Patel, V., R. Araya, M. de Lima, A. Ludermir, and C. Todd. 1999. "Women, Relationships of Trauma to Symptoms of Depression and Post- Poverty and Common Mental Disorders in Four Restructuring Traumatic Stress Disorder among Cambodian Survivors of Mass Societies." Social Science and Medicine 49 (11): 1461­71. Violence." British Journal of Psychiatry 173: 482­88. Patel, V., F. Gwanzura, E. Simunyu, K. Lloyd, and A. Mann. 1995. "The Mollica, R., K. McInnes, N. Sarajlic, J. Lavelle, I. Sarajlic, and M. P. Phenomenology and Explanatory Models of Common Mental Massagli. 1999. "Disability Associated with Psychiatric Co-Morbidity Disorders: A Study in Primary Care in Harare, Zimbabwe." Psychology and Health Status in Bosnian Refugees Living in Croatia." Journal of the and Medicine 25 (6): 1191­99. American Medical Association 282 (5): 433­39. Perriens, J. H., M. Mussa, M. K. Luabeya, K. Kayembe, B. Kapita, C. Brown, Molteno, G., C. D. Molteno, G. Finchilescu, and A. R. Dawes. 2001. P. Piot, and R. Janssen. 1992. "Neurological Complications of HIV-1- "Behavioural and Emotional Problems in Children with Intellectual Seropositive Internal Medicine Inpatients in Kinshasa, Zaire." Journal Disability Attending Special Schools in Cape Town, South Africa." of the Acquired Immune Deficiency Syndrome 5 (4): 333­40. Part 6. Journal of Intellectual Disabilities Research 45: 515­20. Rumble, S., L. Swartz, C. Parry, and M. Zwarenstein. 1996. "Prevalence of Morrison, M. F., J. M. Petitto, T. Ten Have, D. R. Gettes, M. S. Chiappini, Psychiatric Morbidity in the Adult Population of a Rural South African A. L. Weber, P. Brinker-Spence, R. M. Bauer, S. D. Douglas, and D. L. VIllage." Psychological Medicine 26: 997­1007. Evans. 2002. "Depressive and Anxiety Disorders in Women with HIV Rutter, M., and D. Quinton. 1984. "Parental Psychiatric Disorder: Effects Infection." American Journal of Psychiatry 159: 789­96. on Children." Psychological Medicine 14 (4): 853­80. Mulugeta, E., M. Kassaye, and Y. Berhan. 1998. "Prevalence and Outcomes Rwegellera, G. G., and C. C. Mambwe. 1977. "Psychiatric Status and the of Sexual Violence among High School Students." Ethiopian Medical Disorders of Thyroid Function. 1. Prevalence of Goiter in a Group of Journal 36 (3): 167­74. Psychiatric Patients." Medical Journal of Zambia 11 (3): 78­83. Muniu, E., M. N. Katsivo, L. W. Mwaura, and M. Amuyunzu. 1994. "Fatal Sartorius N., A. Jablensky, A. Korten, G. Ernberg, M. Anker, J. E. Cooper, Non-Transport Injuries in Nairobi, Kenya." East African Medical and R. Day. 1986. "Early Manifestations and First-Contact Incidence of Journal 71 (6): 346­49. Schizophrenia in Different Cultures. A Preliminary Report on the Initial Evaluation Phase of the WHO Collaborative Study on Murray, C. J. L., and A. D. Lopez. 1996. Global Health Statistics: A Determinants of Outcome of Severe Mental Disorders." Psychological Compendium of Incidence, Prevalence, and Mortality Estimates for over Medicine 16 (4): 909­28. 200 Conditions. Cambridge, MA: Harvard School of Public Health. Sartorius, N., T. B. Ustun, Y. Lecrubier, and H. U. Wittchen. 1996. ------. 1997a. "Alternative Projections of Mortality and Disability by "Depression Comorbid with Anxiety: Results from the WHO Study on Cause, 1990­2020: Global Burden of Disease Study." Lancet 349: Psychological Disorders in Primary Health Care." British Journal of 1498­1504. Psychiatry 168 (Suppl. 30): 38­43. ------. 1997b. "Global Mortality, Disability, and the Contribution of Risk Schier, E., T. Yecunnoamlack, and T. Tegegne. 1989. "Neuropsychiatric Factors: Global Burden of Disease Study." Lancet 349: 1436­42. Syndromes in Childhood and Adolescence among Patients at a Mzezewa, S., K. Jonsson, M. Aberg, and L. Salemark. 2000. "A Prospective Neuropsychiatric Ambulatory Care and Pediatric Unit of a Teaching Study of Suicidal Burns Admitted to the Harare Burns Unit." Burns Hospital in Ethiopia"(in German). Arztliche Jugendkunde 80 (3): 140­46. 26 (5): 460­64. Schlebusch, L. 1985. "Self-Destructive Behaviour in Adolescents." South Narayan, D. 2000. Voices of the Poor: Can Anyone Hear Us? London: Oxford African Medical Journal 68 (11): 792­95. University Press. Sebit, M. B. 1995. "Neuropsychiatric HIV-1 Infection Study: In Kenya and Narayan, D., R. Chambers, M. K. Shah, and P. Petesch. 2000. Voices of the Zaire Cross-Sectional Phase I and II." Central African Journal of Poor: Crying Out for Change. London: Oxford University Press. Medicine 41 (10): 315­22. Nemeroff, C. B., D. L. Musselman, and D. L. Evans. 1998. "Depression and Setel, P. W., N. Unwin, K. G. M. M. Alberti, and Y. Hemed. 2000. "Cause- Cardiac Disease." Depression and Anxiety 8 (Suppl. 1): 71­79. Specific Adult Mortality: Evidence from Community-Based Nhiwatiwa, S., V. Patel, and W. Acuda. 1998. "Predicting Post-Natal Mental Surveillance--Selected Sites, Tanzania, 1992­1998." Morbidity and Disorder with a Screening Questionnaire: A Prospective Cohort Study Mortality Weekly Report 49: 416­19. from Zimbabwe." Journal of Epidemiology and Community Health 52 (4): Spire, B., S. Duran, M. Souville, C. Leport, F. Raffi, J. P. Moatti, and APRO- 262­66. CO Cohort Study Group. 2002. "Adherence to Highly Active Mental Health and the Abuse of Alcohol and Controlled Substances | 349 Antiretroviral Therapies (HAART) in HIV-Infected Patients: From a ------. 1999b. Global Status Report on Alcohol. Geneva: WHO. Predictive to a Dynamic Approach." Social Science and Medicine 54: ------. 2001. Atlas: Country Profiles on Mental Health Resources. Geneva: 1481­96. WHO. Stewart-Brown, S. 1998."Emotional Well-Being and Its Relation to Health. ------. 2002a. World Health Report 2002--Reducing Risks, Promoting Physical Disease May Well Result from Emotional Distress." British Healthy Life. Geneva: WHO. Medical Journal 317: 1608­9. ------. 2002b. World Report on Violence. Geneva: WHO. Strauss, P. R., C. A. Gagiano, P. H. J. J. van Rensburg, K. J. de Wet, and H. J. ------. 2005. Mental Health Atlas: Revised Edition. Geneva: WHO. Strauss. 1995. "Identification of Depression in a Rural General Practice." South African Medical Journal 85 (8): 755­59. WHO, NEPAD (New Partnership for Africa's Development), and ACOSHED (African Council for Sustainable Health Development). Turner, B. J., C. Laine, L. Cosler, and W. W. Hauck. 2003. "Relationship of 2005. Taking the Human Resources for Health Agenda Forward at the Gender, Depression, and Health Care Delivery with Antiretroviral Country Level in Africa. Report on the Consultative Meeting, Adherence in HIV-Infected Drug Users." Journal of General Internal Brazzaville, Republic of Congo, July. Medicine 18: 248­57. Wilk, C. M., and P. Bolton. 2002. "Local Perceptions of the Mental Health Ustun, T. B., J. L. Ayuso-Mateos, S. Chatterji, C. Mathers, and C. J. L. Effects of the Uganda Acquired Immunodeficiency Syndrome Murray. 2004. "Global Burden of Depressive Disorders in the Year Epidemic." Journal of Nervous and Mental Disease 190 (6): 394­97. 2000." British Journal of Psychiatry 184: 386­92. Wilson, D. A., and P. J. Wormald. 1995. "Battery Acid--An Agent of Vaz, R. F., E. F. Mbajiorgu, and S. W. Acuda. 1998. "A Preliminary Study of Attempted Suicide in Black South Africans." South African Medical Stress Levels among First Year Medical Students at the University of Journal 85 (6): 529­31. Zimbabwe." Central African Journal of Medicine 44 (9): 214­19. World Bank. 2001. Core Welfare Indicators Survey: Handbook and CD-ROM. Verrier-Jones, P., F. D. Pascoe, L. S. Gillis, and J. B. King. 1978. "The First Africa Operational Quality and Knowledge Services. Washington, 100 Patients in the Valkenberg Psychogeriatric Assessment Unit." South DC: World Bank. African Medical Journal 54 (3): 113­15. Yegomawork, G., N. Deyessa, Y. Berhane, M. Ellsberg, M. Emmelin, M. Weiss, E. L., J. G. Longhurst, and C. M. Mazure. 1999. "Childhood Sexual Ashenafi, A. Alem, et al. 2003. "Women's Health and Life Events Study Abuse as a Risk Factor for Depression in Women: Psychosocial and in Rural Ethiopia." Special issue, Ethiopian Journal of Health Neurobiological Correlates." American Journal of Psychiatry 156: Development 17. 816­28. Zellweger, M. J., R. H. Osterwalder, W. Langewitz, and M. E. Pfisterer. 2004. WHO (World Health Organization). 1999a. Facts and Figures about "Coronary Artery Disease and Depression." European Heart Journal Suicide. Geneva: WHO. 25 (1): 3­9. 350 | Florence K. Baingana, Atalay Alem, and Rachel Jenkins Chapter 23 Neurological Disorders Donald Silberberg and Elly Katabira Neurological disorders are increasingly prevalent in Sub- treatable. For example, most developmental disorders and Saharan Africa. The factors that are producing this increased many strokes are preventable. Epilepsy, a common problem, burden include malnutrition, adverse perinatal conditions, is potentially treatable with currently available low-cost malaria, the human immunodeficiency virus and the medications. Disorders such as HIV/AIDS and the dementia acquired immune deficiency syndrome (HIV/AIDS) and that often accompanies HIV infection are currently untreat- other causes of encephalitis and meningitis, demographic able, but their prevalence can be drastically reduced with transitions, increased vehicular traffic, and persistent antiretroviral therapy intervention (Sacktor 2002). Although regional conflicts. Leading neurological disorders include few epidemiological studies have been carried out in Sub- cerebral palsy, mental retardation and other developmental Saharan Africa, it is clear that some disorders of the nervous disorders, epilepsy, peripheral neuropathy, stroke, and, system are more prevalent in this World Bank region than increasingly, the nervous system complications of HIV/ elsewhere in the world. Examples of such overrepresentation AIDS, trauma, and alcohol abuse. The disabling rather than include epilepsy, stroke in younger individuals, and neuro- fatal nature of many neurological disorders, the stigma asso- logical complications related to HIV infection. ciated with brain disorders, and the enormous difficulty in Cultural and religious issues and beliefs are important in gathering epidemiologic data have resulted in their being Sub-Saharan Africa. They influence the value placed by soci- underreported and neglected in Sub-Saharan Africa. This ety on neurological health, the presentation of symptoms, neglect represents an unfortunate paradox, since neurological illness behavior, access to services, pathways through care, (and psychiatric) disorders make up at least 25 percent of the way individuals and families manage illness, the way the the global burden of disease and are responsible for an even community responds to illness, the degree of acceptance and greater proportion of persons living with disability. support--and stigma and discrimination--experienced by Among the hundreds of specific disorders, some common the person with neurological illness. Because of this, cultural and some uncommon, many are potentially preventable or and other contextual factors are important considerations in 351 developing locally appropriate health care policies, pro- 74 to 80 percent for manner of death, and 71 to 82 percent grams, and services. for mechanism of death. The maintenance cost of this sys- tem is estimated to be R 8.00 (US$1.00) per case registered (Gill et al. 2001). MEASUREMENT AND DATA SOURCES Another useful tool for accumulating data is period prevalence studies, such as those that can be registered at a One of the most challenging public health problems in Africa single health facility. These are relatively inexpensive stud- is data collection. Although community surveys date from ies to conduct, and they are informative regarding outpa- nearly 100 years ago, it is only in the last four decades that tient or hospital use at a particular point of service these and other epidemiological methods have provided (Birbeck 2001). adequate diagnostic information based on standardized Newer epidemiological methods, such as capture-recapture methods of assessment that permit the comparison of methodology and log linear modeling, do not seem to have research from different locations and from different levels. been introduced into investigations of neurologic disease in Major challenges to epidemiological research are the Sub-Saharan Africa at this time. The basic problems of data limited capacity of international classification systems for acquisition must first be solved, including designing valid coding disorders and the noninclusion of neurological screening instruments that will allow population-based disorders as separate categories in health management screening. However, the International Clinical Epidemiology information systems. Additionally, systems and instruments Network has fostered the development of clinical epidemio- for epidemiological research of neurological disorders have logy research units in Cameroon, Ethiopia, and Uganda, each not been standardized and validated for use in all areas of of which seeks to apply appropriate methods to their country's Sub-Saharan Africa. The existing capacity and resources to challenges. carry out comprehensive and scientifically sound communi- A PubMed search carried out with the key words "neuro- ty assessments are limited. Many studies have used hospital- logical disorders," and "Sub-Saharan Africa" also confirmed based data, even though many of the patients do not have that few epidemiological studies have been carried out and access to or knowledge of the availability of these services. published in peer-reviewed journals. The best data sources Further challenges to epidemiological research of neuro- are from South Africa and Nigeria and a few from Kenya and logical disorders in Sub-Saharan Africa are the often unreli- Zambia, whereas others, such as those from Uganda, are very able health facility records, the noninclusion of neurolog- old. Additional data are potentially available in unpublished ical disorders as separate items in the health management surveys, particularly in Francophone Africa. information system, and the paucity of cross-sectional studies. No disease surveillance system and no censuses, reg- istries, or other administrative data include neurological EPIDEMIOLOGY disorders. Few of the epidemiological studies of developmental dis- In a unique study, Jelsma and colleagues (2002) conducted orders, Parkinson's disease, stroke, peripheral neuropathies, house-to-house screening visits of 10,839 residents in a dementia, or other disorders carried out in Sub-Saharan "high density suburb" of Harare, Zimbabwe. Visits were Africa have been published in peer-reviewed journals. followed up by medical examination and interview of those Similarly, few studies have been carried out that specifically identified as having a functional limitation. The rate of quantify the neurological complications of HIV/AIDS in disability and morbidity was 5.6 percent for the whole sam- Sub-Saharan Africa in spite of the overwhelming burden of ple. Headaches, including migraine, were the most common the disease in the region. problem. These were followed by back pain, hypertension, In 1998 South Africa began pilot operation of its and osteoarthritis. HIV/AIDS was the fifth most common National Non-Natural Mortality Surveillance System, a use- condition. Depression, based on responses to a screening ful method of accumulating data. "Nonnatural mortality" tool, was evident in one-third of the subjects. Common refers to deaths from homicide, suicide, accidents, and activity limitations included difficulty with the performance undetermined causes. Evaluation carried out in 2001 found of housework and with walking. HIV/AIDS resulted in the sensitivity to range from 65 to 95 percent for manner of most severe activity limitation, in that cognitive functions death, with a positive predictive value that ranged from were also affected. 352 | Donald Silberberg and Elly Katabira Developmental Disorders in Kenya (Feksi et al. 1991). The WHO publication Epilepsy in the African Region provides a more complete account of Few data exist on developmental disorders, other than such epidemiological studies (WHO 2004). reports as that on the pediatric neurology clinic in The Furthermore, epilepsy is a much more frequent cause of Gambia, which indicated that developmental delay, speech death in Sub-Saharan Africa than in wealthy countries. A disorders, and learning difficulties were present among community-based study done in Ethiopia revealed that 128 children during a six-month period (Burton and Allen 6.3 percent of people with epilepsy had died over a two- 2003). A three-year study of a child neurology clinic in year period, and one-third in 20 years (Tekle-Haimanot, Ibadan, Nigeria, found that cerebral palsy accounted for Forsgren, and Ekstedt 1997). In Africa, epilepsy mortality is 16.2 percent of new referrals. Sixty-three percent of these primarily due to status epilepticus, falls, drowning, and cases were judged to have had potentially preventable causes, burns in addition to the sudden death that is known not including intracranial infections (Nottidge and Okogbo throughout the world. 1991). Anecdotal evidence suggests that the prevalence and incidence of mental retardation, cerebral palsy, and other serious developmental disturbances are much higher in Infections Sub-Saharan Africa than in wealthy countries.Studies carried Few studies from Sub-Saharan Africa document the impact out in other low-income regions, which show rates up to of the many infections on neurological morbidity and 24 times higher than in similar studies carried out in Europe mortality. Bacterial meningitis, particularly that due to or North America, support this probability (Durkin 2002). pneumococcal and meningococcal organisms, is still com- mon. The neurological outcomes of these readily treatable Epilepsy infections depend on the availability of and access to health Published figures for prevalence of epilepsy in Sub-Saharan services, which differ tremendously across the region. The Africa range from 2.2 to 58.0 per 1,000 people. Rates are situation is made worse where populations are displaced higher in rural areas, reflecting the higher prevalence of by conflicts. Epidemics of meningococcal disease are partic- predisposing factors, such as higher rates of birth trauma ularly frequent in the "meningococcal belt," which extends and repeated malaria or other parasitic infections. from Guinea to the Sudan and northern Uganda and Hospital-based studies underestimate the prevalence of where, every 5 to 10 years, up to 100,000 cases occur. Viral epilepsy at the community level by at least 30 percent. For encephalitis is on the increase, particularly where the preva- comparison, the available community-based prevalence lence of HIV infection is high as well. data on epilepsy for industrial countries range from 3.3 per The HIV virus enters the nervous system within hours of 1,000 people in England to 6.6 per 1,000 in the United an individual's becoming infected. Acute inflammatory States (IOM 2001, p. 184). demyelinating polyneuropathy (Guillain-Barré syndrome), Community-based surveys are the best way to evaluate which can cause paralysis leading to death from respiratory the magnitude and distribution of epilepsy (IOM 2001 failure, often accompanies this initial HIV infection and p. 186). As one of the few published examples, Osuntokun occurs with greater frequency in those infected with et al. (1978) used an instrument developed for the World HIV/AIDS than in uninfected populations. As HIV/AIDS Health Organization (WHO) by Bruce Schoenberg. At the progresses, all the neurological complications appear to University of Limoges, a questionnaire in French has been occur with the frequencies similar to those that were reported developed for the assessment of epilepsy prevalence in trop- in wealthier countries prior to the advent of treatment with ical countries (Preux 2000). However, the results of such antiretroviral agents. Opportunistic infections of the nerv- surveys are influenced by local cultural factors. Where ous system occur in about 30 to 40 percent of those with epilepsy is regarded as a social stigma, which is common, AIDS (McArthur, Brew, and Nath 2005). An example is the family members hide patients suffering from the condition common cryptococcal meningitis that often develops as from researchers. Because of the difficulties and the expense immune suppression occurs. Other complications include of implementing door-to-door surveys, increased attention toxoplasmosis; herpes zoster; central nervous system (CNS) is being focused on the use of key community informants to lymphoma; several varieties of peripheral neuropathies; identify patients suffering from epilepsy. This strategy was myelopathy, causing paraplegia; strokes; retinal infection, applied successfully in an urban marginal and rural region leading to blindness; and the dementia that develops in Neurological Disorders | 353 approximately 40 percent of those with AIDS. The fact that been attempted in urban and rural Tanzania (Walker et al. most infants and children who acquire HIV from their 2000). Among adults, 5.5 percent of deaths were attributed mothers develop cognitive delay, seizures, and opportunistic to cerebrovascular disease. The yearly age-adjusted rates per nervous system infections is barely recognized in most 100,000 people in the 15 to 64 age group averaged 49 per reviews. 100,000, four times the rates in England and Wales. In South As a result of the HIV pandemic in Sub-Saharan Africa, Africa, stroke accounts for 8 to 10 percent of all reported the neurological complications of HIV infection have deaths and 7.5 percent of deaths among people of prime become major, often overwhelming components of the working age, between 25 and 64 years old (Kahn and overall health burden. These complications are among the Tollman 1999). A prospective community survey in rural most common, and often the most common neurologic South Africa reported that stroke accounted for 25 percent disorders in a population. This is particularly true in East, of all noncommunicable disease, including in many younger central, and southern Africa. Opportunistic infections over- individuals. Stroke was responsible for 5.5 percent of all whelm many facilities, some of which have developed poli- deaths and 10.3 percent in those age 35 to 64 years. Stroke cies that permit only one hospital admission in a patient's ranked second as the cause of death in those age 35 to 64 lifetime. CNS lymphoma, a relatively uncommon tumor years, first in those age 55 to 74 years (11 percent of all ordinarily, is now the most commonly found brain tumor in deaths), and second among those age 75 and older (6 percent many regions. It is important to note that where highly of all deaths) (Kahn and Tollman 1999). In a rural hospital active antiretroviral therapy has become available, the inci- in Zambia, stroke accounted for 9 percent of admissions, dence of these complications among those with HIV has but used 14 percent of the intensive care unit's bed days decreased significantly (Sacktor 2002). (Birbeck 2001). The mortality following stroke was 50 per- Due to the widespread prevalence of (often unrecognized cent, far higher than in wealthy countries, reflecting the lack and untreated) sexually transmitted diseases, the neurologi- of resources for early recognition and access to treatment. cal complications of syphilis remain common. These include meningovascular syphilis, causing strokes; tabes dorsalis, Dementia causing pain and paraparesis; and luetic encephalitis ("general paresis of the insane"), which is rapidly fatal if not Available data indicate that the age-specific prevalence of treated. All these manifestations of infection with Treponema Alzheimer's disease in Nigeria is similar to that among pallidum are preventable by treatment with antibiotics. African Americans (Hendrie et al. 2001). However, age- Malaria rivals AIDS for its impact on the nervous system. related dementia becomes less common as HIV/AIDS low- In addition to the hundreds of thousands of children who ers life expectancy. Those who live long enough to develop die each year from cerebral malaria, many more survive dementia are often regarded culturally as not sick and thus (often repeated attacks) and develop sequelae that have yet do not seek medical care. However, HIV/AIDS has made to be quantified. These include cognitive disorders and dementia a major issue because the associated dementia is epilepsy. Similarly, tuberculous meningitis, and spinal common and often occurs in young people. tuberculosis (TB), leading to spinal cord compression and paraplegia (Pott's disease), are leading causes of death and Movement Disorders disability. Additionally, in many regions leprosy continues to cause deforming peripheral neuropathy, and it may be Parkinson's disease is known to occur in the region, but becoming more common as the result of the redirection of there is little documentation as to its prevalence and impact resources in response to the HIV/AIDS epidemic. on morbidity and mortality. At the neurology clinic in Mulago Hospital, the teaching hospital for the Faculty of Medicine, Makerere University, Uganda, about two to three Stroke new cases are seen every month. However, most of the Stroke is a leading cause of death and disability in Sub- afflicted never seek medical care, particularly in the rural Saharan Africa. To date, most data on mortality have been areas. Sydenham's chorea, now uncommon in wealthy coun- hospital-based, although the majority of stroke deaths in the tries, remains a common problem in Sub-Saharan Africa, region are thought to occur at home (Kahn and Tollman due to the high rates of streptococcal infection and poor 1999). More accurate measures of stroke mortality have access to antibiotics. Additionally, the full spectrum of 354 | Donald Silberberg and Elly Katabira common movement disorders (such as essential tremor) during pregnancy, such as syphilis and rubella; perinatal and uncommon movement disorders (such as chorea gravi- conditions, such as difficult, unassisted deliveries; and darum and focal dystonias) occur. neonatal infections, including tetanus, meningitis, and septicemia. Preventable infections resulting in the extremely high under-five mortality in the region are often the cause Headache of neurological sequelae for those that survive. The frequency From the limited available studies in the region, headache is and adequacy of care for epilepsy and cerebral palsy can be one of the most common neurological complaints encoun- used as indicators of the available quality of the health care tered by health workers. Matuja (1991), at Muhimbili systems. Medical Center, Dar es Salaam, Tanzania, found that recur- In children, febrile convulsions are reported in every rent headaches accounted for 20.6 percent of all new refer- African health structure as a major cause of seizures. rals to the adult neurology clinic over a period of two years. Prevalent causes of fever are similar across the region and Thirty-four percent of the individuals had migraine, 27 per- include malaria, pneumonia, and measles. Other infections, cent had psychogenic (mostly anxiety and depression) such as meningitis, encephalitis, and bacterial septicemia, disorders, and 13 percent had posttraumatic headache may directly affect the brain and cause epilepsy. Other (Matuja 1991). Similarly, migraine accounted for 5.7 per- etiologies of epilepsy include such obvious causes as head cent of all new patients seen in the child neurology clinic at injury, parasitic infestations of the nervous system (such as Estate Specialist Hospital, in Kano, Nigeria (Okogbo 1991). cysticercosis), congenital CNS abnormalities, tumors, and Similarly, data from Zimbabwe indicated that headaches vascular and metabolic disorders. were the most common problem encountered in a large A review of the neurological disorders seen at the pedi- door-to-door survey (Jelsma et al. 2002). These prevalence atric neurology clinic of the University of Nigeria Teaching rates for headache do not differ significantly from those in Hospital in Enugu revealed that perinatal problems, such as the United States and Europe. birth asphyxia, severe neonatal jaundice, and infections, were the most common etiological factors identified (Izuora and Iloeje 1989). Trauma Studies from Côte d'Ivoire, Nigeria, and Zimbabwe found hypertension to be the main risk factor for both In some countries in Sub-Saharan Africa, trauma rivals ischemic and hemorrhagic stroke (Matenga 1997; Walker infections and vascular disorders as the most important 1994). Hypertension is an increasingly important public cause of neurological disease. Moreover, increasing motor- health problem in African countries, where it may affect up ized transportation, violence, and persistent armed con- to 10 percent of the population and contributes to coronary flicts in the region are rapidly changing the pattern of this heart disease, as well as to hemorrhagic and thrombotic trauma. Head trauma, if not immediately fatal, often leads stroke. The condition frequently goes unrecognized, how- to cognitive impairment or epilepsy or both. Spinal cord ever, in part because many African health care providers trauma often produces quadriparesis, paraparesis, or paral- lack reliable equipment for measuring blood pressure ysis. In the absence of skilled rehabilitation services, the (Birbeck 2000). In addition, the limited and erratic supply average life span following significant spinal cord injury is of and access to appropriate drugs for the management of no more than several years. hypertension contributes to the related high morbidity and mortality. Sickle-cell disease is a major cause of stroke among children and young adults in West, East, and central ETIOLOGY AND DETERMINANTS Africa, affecting at least 1 percent of those with sickle-cell disease per year. An analysis of 320 adult stroke patients in The major risk factors for neurological disorders can be Durban, South Africa, revealed that HIV/AIDS, tuberculo- classed as genetic factors; nutritional deficiencies; infection; sis, cysticercosis, and syphilis were the most common causes exposure to environmental toxins; prenatal, perinatal, and of stroke in the young adult age groups. Additionally, neonatal factors; poverty; and trauma. emboli from streptococcal infection­induced cardiac dis- In Sub-Saharan Africa, by far the leading causes of neu- ease (rheumatic heart disease) are a common cause of rological disorders are preventable. These include infections stroke in young adults. Neurological Disorders | 355 CONSEQUENCES OF NEUROLOGICAL DISORDERS Osuntokun (1975) noted that because of the gross short- age of "Western-trained" health personnel, Nigerians fre- Neurological disorders in Sub-Saharan Africa impose a quently turn to traditional native medicine and native significant burden on the family and community, as well as herbalists (as do many in other countries in Sub-Saharan on the affected individual. Some disorders, such as epilepsy, Africa). Although the training of these native doctors is are well recognized but are not socially and culturally arduous and lasts an average of 8 to 10 years, their short- acceptable. The enormous stigma attached to epilepsy often comings often interfere with administration of effective leads to the patient's being denied access to proper care. treatments in community clinics. These doctors are excel- Lack of care then leads to severe complications, social isola- lent psychotherapists and maintain good relationships with tion, and early death. Other disorders, such as dementia, their patients. Yoruba native doctors recognize several paraplegia, and stroke, put undue stress on the caring family classical neurological diseases, including epilepsy, cere- because institutional or community care support is limited brovascular disease, fever and headache, migraine, and or nonexistent. This stress leads to further neglect of the ataxic neuropathies. However, the scientific basis of their afflicted person and eventually his or her premature death. pharmacotherapeutics is for the moment mostly unknown. Protein-calorie and micronutrient malnutrition contribute Their practice is thus partially based on the very large body to impaired cognitive development, which compromises the of knowledge of herbal medicines in Africa that has yet to be future productivity of a nation's workforce. Although explored scientifically. the more complete epidemiology and cost analyses have yet The role of the traditional healer is sometimes problem- to be done, there seems little doubt that neurological disor- atic. People who would without hesitation go to their local ders impede economic development in many Sub-Saharan health care center with a severe cough, fever, or burn may Africa countries. seek care for epilepsy through their local healer for years, never revealing their problem to the health care workers treating their other problems, and thus not receiving appro- RESOURCE ISSUES priate treatment. In a personal communication, Gretchen Birbeck, an asso- Virtually all of the excess morbidity and mortality that ciate professor of neurology and epidemiology at Michigan occur as a consequence of neurological disorders in Sub- State University, had this to say, based on her extensive expe- Saharan Africa result from scarcities of resources. Two rience in Zambia: major subsets of resource scarcities that are particularly It is important also to recognize that traditional healers important to neurological health and disease are the num- may fail to recognize or refer urgent treatable conditions bers of specifically trained health care workers, and the diag- and that their therapies can have adverse effects. Children nostic (and therapeutic) resources available to them. whose parents attribute their malaria-associated seizures to supernatural causes suffer from higher parasite counts Human Resources and require longer lengths of hospital stays than their peers, likely associated with initially seeking care from Countries in Sub-Saharan Africa do not have sufficient local healers. Until traditional healers can become more qualified staff in the clinical neurosciences (neurologists, connected to the biomedical healthcare system and/or neurosurgeons, psychiatrists). Except for South Africa, the the public becomes a great deal more educated, tradi- mean ratios for countries that have these medical specialists tional healers will continue to care exclusively for patients are 1 neurologist for 1 million to 2.8 million people (versus who would benefit greatly from medications and treat- 4 per 100,000 in Europe); 1 psychiatrist for 900,000 people ments that are routinely available in most Sub-Saharan (versus 9 per 100,000 in Europe); and 1 neurosurgeon for Africa hospitals. 2 million to 6 million people (versus 1 per 100,000 in Europe). Most of the clinical neuroscience services are located in the Physicians are too scarce to provide routine outpatient capital cities, often the largest urban areas, where the profes- care in most settings. The nurses and community health sionals also often lecture at the medical schools. Neurology workers who are the frontline providers receive little or patients often must travel long distances to consult with a no training in how to diagnose and treat the common neu- doctor in the city. rologic conditions that present to their clinic every day 356 | Donald Silberberg and Elly Katabira (Birbeck and Munsat 2002). Their inability to respond to irreversible. On ethical grounds alone, prevention is always neurologic complaints reinforces the patients' inclination to preferable to treatment or rehabilitation. In most instances seek care from traditional healers. In addition, as neurology prevention is also more cost-effective than treatment. Many has evolved into a subspecialty separate from general potentially catastrophic disorders are now preventable. medicine, fewer generalists receive neurologic training. Preventive measures include immunization against tetanus, The adequate supply of physicians in wealthier countries tuberculosis, measles, rubella, and polio, all of which pro- has resulted in greater use of neurologic consultations. duce neurological dysfunction as their primary route of However, where neurologists are lacking, as in Sub-Saharan attack, or as a very common complication; early and effec- Africa, this segregation of neurology from general medicine tive management of childhood fevers to prevent febrile leaves a vacuum. Because it may take many years to increase convulsions; and the use of zinc, folic acid, and iron supple- the number of neurologists, better training for primary care mentation and fortification to facilitate normal brain devel- providers is essential. opment. Prevention of HIV transmission through effective health education strategies and programs to prevent moth- Diagnostic Facilities: Instrumentation er to child transmission are critical. Short-course antiretro- viral prophylaxis regimens are becoming widely used in The most important diagnostic measures are through inter- such programs in Sub-Saharan Africa and, in conjunction views and by the usual clinical examination, looking for with supplemental feeding, are likely to significantly reduce signs and symptoms of disturbed neurologic functions. This HIV transmission. The situation will improve still further can be accomplished in any health structure. However, accu- when antiretroviral therapy becomes universal in the region. rate diagnosis and management often requires diagnostic Safe motherhood initiatives also greatly reduce the impact means that are usually found in secondary and tertiary of prenatal and perinatal risk factors. health care facilities. In Sub-Saharan Africa, with the excep- The potential for prevention of stroke is illustrated by the tion of South Africa, the specialized diagnostic tools for fact that in wealthy countries the incidence of stroke has brain disorders, necessary for such purposes, consist of been reduced by 25 to 40 percent, largely because of early 79 EEG machines, 65 CT scanners, and 9 MRIs. In contrast, recognition and treatment of hypertension and the reduc- South Africa has 60 EEG machines, 214 CT scanners, and tion in tobacco use (Rothwell et al. 2004; Whisnant 1984). 46 MRIs (WHO 2004). Other preventive strategies include increased exercise, To compound the problem, this equipment is often badly adopting healthy eating habits, and regular monitoring of maintained or out of order. Further, most of the patients cholesterol. Prevention programs, although highly cost- referred to capital city hospitals cannot afford the price and effective, have not been undertaken in Sub-Saharan Africa. therefore cannot benefit from the technological progress. The few neurological surgeons practicing in Sub-Saharan Africa not only face these shortages in diagnostic equipment Treatment but must cope with a scarcity of needed surgical instruments The WHO recommends phenobarbital as the first-line as well as with a general paucity of the institutional resources drug for the treatment of partial and generalized tonico- needed to carry out major procedures successfully. clonic epilepsy in developing countries. The use of this old and simple drug is encouraged because its efficacy for a wide range of seizure types and its low cost make it suit- INTERVENTIONS: WHAT HAS WORKED? able for use in primary health care in developing countries. Despite the problems enumerated above, many avenues are The majority of affected individuals can be treated suc- open to reduce the impact of neurological disorders in Sub- cessfully with phenobarbital at a cost of about US$5 per Saharan Africa. The potential advances start with preven- year, thereby avoiding disability and the risk of premature tion, and proceed to treatment and rehabilitation. death. Phenytoin, which may be more effective for some individuals, has few side effects and cost US$20.59 in Sub- Saharan Africa in 2000 (WHO 2001). The publication Primary Prevention Epilepsy in the African Region (WHO 2004) provides a Prevention is critical in reducing the incidence of brain more complete account of the availability and cost of disorders. Impairment caused by these disorders is often treatment, by country. Neurological Disorders | 357 The current global effort to make antiretroviral drugs environment. Socioeconomic determinants of the outcome readily available and affordable to a much larger infected measures that could best reflect optimal care for epilepsy population in the region will reduce morbidity and mortal- patients should be identified. ity related to neurological complications due to HIV/AIDS. Most of the same observations can be made about stroke prevention and treatment, largely unrecognized major problems in Sub-Saharan Africa, and the many other disor- A FRAMEWORK FOR RESEARCH ders that affect the nervous system. For example, the direct and indirect costs of stroke are enormous, including loss of Although disability-adjusted life years (DALYs) have been wage-earning capacity and the need for the family's provi- computed for some of the specific causes of developmental sion of care for survivors. disability, such as meningitis and iodine deficiency,1 these Evaluations of cost-effectiveness of neurological inter- figures do not convey the full proportion of cases within a ventions have not been extensively carried out in the region. given category of disorder that result in early and lifelong They depend on the actual costs of specific drugs and other disability or death (see IOM 2001). DALY estimates are not treatments, which vary among countries and will decrease as currently available for developmental disability as a whole. newer drugs come off patent. What is needed before useful DALY or other measures of impact can be calculated for developmental disability is accurate and up-to-date information from Sub-Saharan POLICY ISSUES Africa countries on the prevalence and impacts of long-term functional limitations originating early in life, as a result of Among brain disorders, epilepsy stands out, not only both known and unknown causes. because of its high prevalence and incidence rates and the Stigmatization of those with epilepsy, and often their potential for successful treatment, but because of the myths families, is widespread. Some cultural beliefs, such as the and beliefs attached to it in various cultures and the result- fear that epilepsy is contagious and that the individual is ing impact on the individual, family, and the community. possessed of demons, confound epidemiological studies, Epilepsy commonly attacks children and young adults in the prevent treatment, and lead to personal and economic dis- most productive years of their lives and frequently leads to aster for many affected individuals and their families. unemployment, which confounds not only the problems of Because stigma and other cultural beliefs are locally grounded, the afflicted but often the family that relies on their financial they require local study in order to develop ways in which to support. deal with them constructively. To implement effective programs for the community- The full picture of the epidemiology of epilepsy remains level detection and treatment of patients with epilepsy, sketchy. In addition to descriptive studies, analytical governments and health authorities of developing countries research is needed to ascertain specific risk factors for the must first issue clear policies articulating measures for the development of the disease with an emphasis on its relation identification and treatment of patients with epilepsy. to infectious disorders, such as pneumonia, meningitis, Essential components of any strategy are a continuous drug and tetanus; its links to the common causes of under-five supply and means of distribution, prevention programs, and neonatal mortality; and its links to birth trauma. social programs and stigma mitigation strategies to increase Environmental and genetic causes should be explored in employment and improve the general welfare of individuals various geographic locations and among different ethnic suffering from the disease, training of health care personnel, groups. Information gathered through such efforts would be and methods for surveillance and routine data collection. critical to the formulation of effective preventive strategies. Increasing the clinical and research professional capacity Most econometric studies of epilepsy in developing in developing countries is key to developing sustainable, countries await implementation. Cost-of-illness, cost- locally appropriate programs. Approaches to addressing effectiveness, cost-minimization, and cost-benefit analyses these deficits include the following: should be undertaken to provide government officials and funding agencies with the necessary economic perspective. · Capacity building through training via on-site education, Furthermore, quality of life should be measured for epileptic exchange programs, and distance learning using modern patients receiving treatment in an economically constrained information technology 358 | Donald Silberberg and Elly Katabira · Development of local networks that link centers with Hendrie, H. C., A. Ogunnii, K. S. Hall, O. Baiyewu, F. W. Unverzagt, O. Gureje, S. Gao, et al. 2001. "The Incidence of Dementia and AD in expertise to their surrounding communities Two Communities: Yoruba Residing in Ibadan, Nigeria, and African · Development of regional networks of centers with Americans Residing in Indianapolis, USA." Journal of the American expertise Medical Association 285: 739­47. · Development of educational programs to enable primary IOM (Institute of Medicine). 2001. Neurological, Psychiatric, and Developmental Disorders: Meeting the Challenges in the Developing care health workers to recognize, and when appropriate, World. Report. Washington, DC: Academy Press. treat brain disorders Izuora, G. I., and S. O. Iloeje. 1989. "A Review of the Neurological · Introduction of locally relevant basic, clinical, and health Disorders Seen at the Paediatric Neurology Clinic of the University of Nigeria Teaching Hospital, Enugu." Annals of Tropical Paediatrics 9 (4): care policy research that will lead countries to choose to 185­90. increase capacity. Jelsma, J., J. Mielke, G. Powell, W. De Weerdt, and P. De Cock. 2002. "Disability in an Urban Black Community in Zimbabwe." Disability and Rehabilitation 24 (16): 851­59. Clinical neuroscientists should be part of the teams that Kahn, K., and S. Tollman. 1999. "Stroke in Rural South Africa-- address the issues of the health care systems in Sub-Saharan Contributing to the Little Known about a Big Problem." South African Africa, so as to be certain that the prevention, recognition, Medical Journal 89 (1): 63­65. treatment, and rehabilitation of brain disorders are Matenga, J. 1997. "Stroke Incidence Rates among Black Residents of Harare--A Prospective Community-Based Study." South African adequately addressed. Medical Journal 87: 606­9. Matuja, W. B. 1991. "Headache: Patterns and Features as Experienced in a Neurology Clinic in Tanzania." East African Medical Journal 68 (12): 935­43. NOTE McArthur, J., B. Brew, and A. Nath. 2005. "Neurological Complications of HIV Infection." Lancet Neurology 4: 543­55. 1. The most recent DALY figures in low- and middle-income countries Nottidge, V. A., and M. E. Okogbo. 1991. "Cerebral Palsy in Ibadan, for risk factors for neurological disorders discussed in this chapter include Nigeria." Developmental Medicine Child Neurology 33: 241­45. HIV/AIDS, 5.5 percent; polio, 0 percent; measles, 2.4 percent; tetanus, 1.0 percent; meningitis, 0.4 percent; malaria, 3.1 percent; Japanese Okogbo, M. E. 1991. "Migraine in Nigerian Children--A Study of 51 encephalitis, 0 percent; trachoma, 0.1 percent; protein-energy malnutri- Patients." Headache 31 (10): 673­76. tion, 1.2 percent; iodine deficiency, 0.1 percent; Vitamin A deficiency, Osuntokun, B. 1975. "The Traditional Basis of Neuropsychiatric Practice 0.2 percent; anemias, 1.9 percent; road traffic accidents, 2.7 percent; homi- among the Yorubas of Nigeria," Tropical Geographic Medicine 27: cide and violence, 1.6 percent; war, 1.7 percent. 422­30. Osuntokun, B. O. 1978. "Epilepsy in Africa. Epidemiology of Epilepsy in Developing Countries in Africa." Tropical Geographic Medicine 30: 23­32. REFERENCES Preux, P. M. 2000. "Questionnaire des investigations de l'épilepsie dans les pays tropicales." Bulletin Société Pathologique Exotique 93: 276­78. Birbeck, G. 2000. "Barriers to Care for Patients with Neurologic Disease in Rothwell, P., A. Coull, M. Giles, S. Howard, L. Silver, L. Bull, S. Gutnikov, Rural Zambia." Archives of Neurology 57 (3): 414­17. et al. 2004. "Change in Stroke Incidence, Mortality, Case-Fatality, Severity, and Risk Factors in Oxfordshire, UK from 1981 to 2004 ------. 2001. "Neurologic Disease in a Rural Zambian Hospital." Tropical (Oxford Vascular Study)." Lancet 363: 1925­33. Doctor 31: 82­86. Tekle-Haimanot, R., L. Forsgren, and J. Ekstedt. 1997. "Incidence of Birbeck, G., and T. Munsat. 2002. "Neurologic Services in Sub-Saharan Epilepsy in Rural Central Ethiopia." Epilepsia 38: 541­46. Africa: A Case Study among Zambian Primary Healthcare Workers." Journal of the Neurological Sciences 200: 75­78. Sacktor, N. 2002. "The Epidemiology of Human Immunodeficiency Virus-Associated Neurological Disease in the Era of Highly Active Burton, K. J., and S. Allen. 2003. "A Review of Neurological Disorders Antiretroviral Therapy." Journal of Neurovirology 8 (Suppl. 2):115­21. Presenting at a Paediatric Neurology Clinic and Response to Anticonvulsant Therapy in Gambian Children." Annals of Tropical Walker, R. 1994. "Hypertension and Stroke in Sub-Saharan Africa." Paediatrics: International Child Health 23 (2): 139­43. Transactions of the Royal Society of Tropical Medicine and Hygiene 88: 609­11. Durkin, M. 2002. "The Epidemiology of Developmental Disabilities in Low-Income Countries." Mental Retardation and Developmental Walker, R. W., D. G. McLarty, H. M. Kitange, D. Whiting, G. Masuki, D. M. Disabilities Research Reviews 8: 206­11. Mtasiwa, H. Machibya, N. Unwin, and K. G. Alberti. 2000. "Stroke Mortality in Urban and Rural Tanzania-Adult Morbidity and Feksi, A., J. Kaamugisha, J. Sander, S. Gatiti, and S. Shorvon. 1991. Mortality Project." Lancet 355: 1684­87. "Comprehensive Primary Health Care Antiepileptic Drug Treatment Programme in Rural and Semi-Urban Kenya." Lancet 337: 406­9. Whisnant, J. P. 1984. "The Decline of Stroke." Stroke 15: 160­68. Gill, G., B. Scott, N. Beeching, D. Wilkinson, and A. Ismail. 2001. WHO (World Health Organization). 2001. Atlas Country Profiles on "Enumeration of Non-Communicable Disease in Rural South Africa Mental Health Resources. Geneva: WHO. http://www.who.int/ by Electronic Data Linkage and Capture-Recapture Techniques." mental_health/management/epilepsy_in_African-region.pdf. Tropical Medicine and International Health 6: 435­41. ------. 2004. Epilepsy in the African Region. Geneva: WHO. Neurological Disorders | 359 Chapter 24 Violence and Injuries Brett Bowman, Mohamed Seedat, Norman Duncan, and Olive Kobusingye Historically, injuries have been understood as inescapable and analytical clarity, but recent evidence points to a cluster realities of everyday life. Increasingly, the timely and accu- of shared risks across intentional and unintentional injuries. rate collection and analysis of data in various parts of the Furthermore, intentionality cannot always be ascertained in world has encouraged a revision of these assumptions. particular circumstances, and violence may indirectly Careful scrutiny of such data has revealed that both inten- contribute to the prevalence of unintentional injuries tional and unintentional injuries are preventable and in (Berger and Mohan 1996). The "intentionality divide" is thus many respects subject to elements of control. established as a useful concept for injury prevention An injury may be defined as "the physical damage that programs, but risk factors appear porous across it. results when a human body is suddenly or briefly subjected Whatever the conceptual ambiguities may be, the WHO to intolerable levels of energy"(Holder et al. 2001, 5). Injuries estimates that injuries constitute 16 percent of the global are traditionally grouped according to two broad categories: burden of disease (WHO 2002a). This translates into intentional and unintentional. Conventionally, intentional 5.8 million injury-related deaths at a rate of 97.9 per 100,000 injuries are comprised of interpersonal violence (spousal worldwide. Injuries further account for between 10 and abuse, child abuse, other assaults), self-inflicted injuries 30 percent of all hospital admissions and render at least (attempted and completed suicides), as well as collective vio- 78 million people disabled each year (Berger and Mohan lence and war-related injuries. Motor vehicle injuries, burns, 1996). In 1998, unintentional injuries accounted for just falls, drownings, and other injury classifications in which under 3.5 million deaths worldwide (WHO 2000b). The intentionality is understood to be absent constitute the burden resulting from unintentional injuries tends to be broad unintentional injuries category. Thus, whereas inten- higher in low- to middle-income countries (LMICs) and tional injuries are associated with violence, as defined by the communities. For instance, 87.9 percent of all road traffic World Health Organization (WHO), unintentional injuries deaths, and 88.3 percent of lost disability-adjusted life years are not. Such a distinction may be valuable for conceptual (DALYs) were from LMICs (Mathers et al. 2001). 361 Despite the heavy burden from the human immunodefi- (Smith and Barss 1991) to the effective execution of this ciency virus and the acquired immune deficiency syndrome mandatory recording and the poor or absent record keeping (HIV/AIDS), malaria, and other infectious diseases, injuries among most vulnerable populations and communities, such were still responsible for 19.93 percent of all deaths of those as refugees, displaced persons, and communities afflicted by between the ages of 15 and 59 years, and for nearly one in civil and armed conflict, national mortality data sets make every four deaths of those between 15 and 29 years. Most of for sound starting points from which to describe the preva- these deaths resulted from road traffic injuries, wars, and lence and magnitude of injury in Africa. These figures, interpersonal violence. According to the WHO (2002a), however, represent the tip of the injury iceberg, because road traffic injuries, war, and homicide, respectively, were although nonfatal injuries are estimated to exceed the global the 10th, 11th, and 14th leading causes of mortality in Africa fatal injury profile by as much as 20 times (WHO 2002a), during 2000. surveillance systems for capturing and reporting nonfatal In the year 2000, an estimated 1.6 million people died injury are far less developed than those recording fatal from various forms of violence (WHO 2002a). Although injury (WHO 2002a). Most of the data on injury are violence manifests as a worldwide public health concern, acquired from surveys and specialized studies within rela- the epidemiology of violence indicates that the majority of tively limited population groups. These studies provide a violent deaths occur in LMICs. Less than 10 percent of all patchwork picture of the injury profile of Africa. The violence-related deaths occurred in high-income countries African injury profile provided here is thus made up of a full (HICs) in 2000 (WHO 2002a). The WHO (2002a, 5) defines supplement of smaller studies extracted from the literature violence as "the intentional use of physical force or power, that complement the Africa-specific data subsets gleaned threatened or actual, against oneself, another person, or from information provided by the WHO. against a group or community, that either results in or has a high likelihood of resulting in injury, death, psychological INTENTIONAL INJURIES harm, maldevelopment or deprivation." This chapter covers the incidence, prevalence, and mag- Intentional injuries (violence) resulted in the deaths of some nitude of intentional (violent) and unintentional mortality 311,000 people in Africa in the year 2000. This translates resulting from injuries at the global and continental into a rate of 60.9 deaths per 100,000 people in Africa as a (African) levels. The injury profiles of South Africa and direct result of intentional injury alone. This figure dwarfs Uganda are examined to demonstrate the merits of estab- the unintentional injury mortality rates of both the WHO lishing a country-level injury surveillance system. The chap- European and American regions with rates of 32.0 (WHO ter also includes an overview of the risk factors and determi- 2002a) and 27.7 unintentional injury deaths per 100,000 nants that are associated with intentional and unintentional people, respectively. As indicated earlier, morbidity and dis- injuries resulting in mortality in Africa. In conclusion, ability due to intentional injury have a prevalence that generic recommendations are listed for the control, preven- exceeds mortality by at least 20 times. Intentional injuries tion, and elimination of injuries and violence in Africa, with therefore resulted in the disability or incapacitation of at a view to encouraging injury control and safety promotion least 6.2 million people on the African continent in the year practice and policy. 2000. Data collected and analyzed in some selected African states indicate the dire burden that injury exerts on these countries. In Zimbabwe, injuries were reported to account DATA SOURCES for 15 percent of all deaths for the year 1988 (Zwi et al. 1993). Survey data suggest that the injury contribution to The global, regional, and national burden of nonfatal injury mortality is as significant in both Ghana and Kenya is exceedingly difficult to measure. Perhaps the most reliable (Forjuoh, Zwi, and Romer 1996). indicators of the magnitude of violence and injury are national mortality data sets. The mandatory recording and Homicide certification of deaths in most countries renders mortality information an adequate source from which to calculate the According to the WHO (2002a) over half a million frequency and prevalence of violence and injuries within (520,000) people died as a direct result of homicide at a rate national and global contexts. Despite the various obstacles of 8.8 per 100,000 in the year 2000. More than three-quarters 362 | Brett Bowman, Mohamed Seedat, Norman Duncan, and Olive Kobusingye (77 percent) of these victims were male (WHO 2002a). Recent studies estimate that 310,000 people died as a direct Globally, the highest levels of homicide occurred among result of injury incurred during warfare in 2000 (WHO males 15 to 29 years old, closely followed by those 30 to 2002a). Injuries resulting from collective violence have been 44 years old. The gender disparity in the worldwide distri- concentrated in Sub-Saharan Africa, Latin America, and the bution of homicide is most apparent through a rate com- Caribbean (Jansson and Svanstrom 1999). The highest rates parison. For every female victim of homicide, 3.4 males are of these wartime injuries occurred in Africa, where war- killed per 100,000 people globally. Conversely, more women related deaths numbered 32 per 100,000 people. Africa thus than men are the victims of intimate partner violence, and contributed 167,000, or 53.8 percent, of war-related injury women are more likely than men to die from such violence mortality to the WHO injury data in 2000. Colonial and (NCIPC 2003). Homicide rates vary according to region and postcolonial Africa has been an epicenter for collective vio- income levels (WHO 2002a). Recent results from the South lence over the last 100 years. Indeed, civil strife and political African National Injury Mortality Surveillance System instability have characterized the continent throughout its (NIMSS) indicate that homicide contributed 36 percent modern history. Moreover, the wars in Africa have tended to to all nonnatural injury deaths in that country in 2000 have increasingly long durations. The increased sophistica- (Burrows et al. 2001) and that homicide continues to be the tion of weaponry has changed the nature of global conflicts. leading cause of premature death among South African Examples of wars on the African continent are numerous. males (Bowley, Parmar, and Boffard 2004). The WHO has estimated that the war between Ethiopia and Eritrea toward the concluding years of the twentieth century Suicide resulted in the deaths of tens of thousands of people. Besides the obviously direct mortality consequences for African Suicide was the cause of death for an estimated 815,000 states, wars undermine health and political infrastructures, people worldwide in 2000 (WHO 2002a). The age-adjusted indirectly resulting in many more deaths from outbreaks of rate of 14.5 per 100,000 attests to the global magnitude communicable disease (WHO 2002a). of self-directed violence. Suicide among males is more Many injury studies in the literature focus on the injury frequent than among females, with over 60 percent of all consequences of African war on children (Cliff and suicides being male (WHO 2002a). An analysis of African Noormohamed 1993; Stohl 2002). Others concentrate on and global suicide reveals that there is a proportionate the direct and indirect health outcomes of war for the increase in suicide with age, suicide reaching its peak among continent in general (Elliot 2000; Elliot and Harris 2001; individuals 60 years of age or older. In this age cohort, male Marysse 2003). Although the bulk of lives lost to war suicide rates are twice those of females. Suicide is most pro- injuries in Africa have resulted from armed conflicts in the nounced in the European, Southeast Asian, and Western Democratic Republic of Congo (Marysse 2003), Liberia, and Pacific regions of the WHO database coverage, where sui- Rwanda, the legacy of war continues to contribute to cide rates are 19.1 per 100,000, 12.0 per 100,000, and 20.8 African mortality through, for example, the detonation of per 100,000, respectively. The suicide rate for African males redundant landmines in Mozambique (Elliot 2000; Elliot in 2000 was 6.7 per 100,000, whereas the rate for females was and Harris 2001). The indirect war-mortality figures are 3.1 per 100,000. Africa appears to be the least frequently equally as devastating as the casualties resulting from direct represented region in studies of suicide, but a number of combat. Estimates of the indirect war-mortality figures national, regional, and citywide comparative studies do vary; some studies estimate 2.5 million Congolese deaths as serve to provide a (limited) profile of the African suicide an indirect consequence of the civil strife that characterized injury burden (Burrows et al. 2003; Meel 2003; Schlebusch the Democratic Republic of Congo from 1998 to 2003 and Bosch 2000). (Marysse 2003). War-Related Injury The number of armed conflicts worldwide has grown expo- UNINTENTIONAL INJURIES nentially over the past century. In the 1950s there were no more than 20 armed conflicts in progress across the globe. The magnitude, prevalence, and severity of unintentional This number rose to 50 in the late 1980s and doubled to 100 injury mortality (bar deaths resulting from natural armed conflicts worldwide during the 1990s (Stohl 2002). disasters) are often underpublicized.Such mortality imposes, Violence and Injuries | 363 however, a significant burden on global and African health in which global female rates of death outnumber those of systems. men. This is, however, not the case in Africa, where males are more frequently the victims of burn fatalities. Young children and the elderly are the most vulnerable to burn Traffic-Related Fatal Injuries injuries. This seems to be the case in all forms of burn mor- According to the WHO (Peden, McGee, and Sharma 2002), tality in particular African countries. A study conducted in the overwhelming majority (90 percent) of all fatal road Tanzania (Mbembati, Maseru, and Leshabari 2002) found traffic injuries (RTIs) occurred in LMICs in 2000. In that that children younger than 15 years constitute a particularly year 1.26 million people across the globe died as a direct vulnerable group for domestic burn injuries. Children in the result of these injuries. This figure translates into a rate of same age cohort have been identified as especially vulnerable 20.8 people dying from RTIs per 100,000 worldwide. Again, to burn injuries in South Africa (Lerel 1994). Forjuoh males are at higher risk of dying in this manner than their (1996) confirmed that children were a vulnerable uninten- female counterparts. For every female traffic-related death, tional (and in 5.4 percent of cases, intentional) burns cohort there are at least three male victims (Peden, McGee, and in the Ashanti region of Ghana. The most frequently repre- Sharma 2002). Globally, males in Africa have the second sented external causes of the initial epidemiological investi- highest RTI fatality rate of 35.8 per 100,000, superseded only gation (Forjuoh 1995) included scalds (44 percent), contact by the male RTI fatality rate of 42.4 per 100,000 in Southeast with hot objects (30 percent), and flames (20 percent). Asia. Worldwide, over 50 percent of all fatal RTIs occur Falls accounted for more than 280,000 deaths globally in among young adults between the ages of 15 and 44 years. the year 2000 (Peden, McGee, and Sharma 2002). Almost a This trend is replicated in Africa (Peden, McGee, and quarter of all falls occurred in high-income WHO regions. Sharma 2002). Globally, females over 70 years of age constitute the most A study focused on describing the prevalence and magni- vulnerable group to fall fatalities. This is particularly pro- tude of RTI fatalities in Kenya (Odero, Khayesi, and Heda nounced in Africa, where female fall fatalities in the age 2003) reported that the country has one of the highest road group 80 years and older are approximately twice that of fatality rates in relation to vehicle ownership in the world, males. and a number of South African studies (De Wet 1993; Deaths due to drowning are almost exclusively an injury Peltzer 2003; Venter 1998 ) have confirmed the significant problem of LMICs. Ninety-seven percent of all drownings contribution of RTIs to the country's nonnatural mortality worldwide occur in LMICs, and thus Africa, where the burden. A recent injury epidemiological study conducted in drowning mortality rate is 113.1 per 100,000 people (Peden Tanzania again revealed RTIs to be the leading cause of and McGee 2003), has been identified as a region at risk injury mortality in the areas of Dar es Salaam, Hai, for drownings. Males in Africa have the highest drowning and Morogoro (Moshiro et al. 2000). A study comparing mortality rates in the world with a drowning rate of 19.2 per transport-related injuries in urban and rural settings in 100,000 population. Children are unquestionably most at Ghana found that the "nature of transport injury in both risk for death from drowning; more than 50 percent of all urban and rural areas is fundamentally different from that in global drowning mortalities occur in children up to 14 years developed countries"(Mock, Forjuoh, and Rivara 1999, 367). old. This trend is emphasized in Africa, where the highest This difference is most pronounced in the contribution of rate of drowning fatalities (18.9 per 100,000) occurs among minibus taxis and other public transport vehicle crashes and children between zero and four years (Peden, McGee, and motor vehicle crashes involving pedestrians (Mock, Forjuoh, Sharma 2002). and Rivara 1999) to transport-related injuries in Ghana. THE EXPERIENCES OF SOUTH AFRICA Burns, Falls, and Drowning AND UGANDA Fire-related burn injuries resulted in the death of some 238,000 people in 2000 (Peden, McGee, and Sharma 2002). Both South Africa and Uganda are countries that can provide Ninety-five percent of these occurred in LMICs. Africa had examples of the benefits of developing sound injury meas- the second highest rate of fatal burn injuries in that year. urement systems. Such systems are aimed at establishing or Most notably, fire-related burns are the only cause of injury implementing African context-specific "best-practices" for 364 | Brett Bowman, Mohamed Seedat, Norman Duncan, and Olive Kobusingye the prevention of injury and can be of use in the various per 100,000 is a serious source of concern. The injury death African states. According to collated international data, rate of 160 per 100,000 in Durban further emphasizes the South Africa contributes significantly to the burden of burden of injury on another one of South Africa's coastal injury in Africa. A sound South African injury mortality cities. In the southwest city of East London, the rate of surveillance system could, however, bias the injury-mortality injury death was 203 per 100,000 for the year 2000. contribution of the country to the continent. Nonetheless, a Transport-related deaths and homicide collectively socioeconomic profile that describes significant disparities account for the majority of deaths in the urban centers of between rich and poor measured by a Gini coefficient of South Africa (Sukhai and Matzopoulos 2002); East London 0.58 (Nattrass and Seekings 2001) and a complex political showed a homicide rate of 100 per 100,000 in 2001 and history make South Africa a good country to analyze as an Cape Town revealed a transport-related fatality crude rate of injury control and prevention case example. The South 42 per 100,000 in 2001 (Sukhai and Matzopoulos 2002). African NIMSS, which captures approximately 34 percent of Although nontransport unintentional injury deaths are all nonnatural deaths nationally, reported 18,876 people to overshadowed by violent and transport-related causes of have died as a direct result of injuries in 2000 (Matzopoulos death in South Africa, their contribution to the injury 2002). Homicide was the leading manner of death (44.5 per- burden of the country is significant; in 2001 alone, East cent), followed by transport-related fatalities (34.5 percent) London had an unintentional, nontransport-related injury and suicide (9.4 percent). Figures extracted from the 2001 mortality rate as high as 27 per 100,000. NIMSS annual report reveal no significant decrease in the Focusing on cities that yield data like those presented magnitude and prevalence of mortality due to injury in above have proved of use in injury prevention and safety South Africa (Matzopoulos 2002). Again these figures repre- promotion in South Africa for many reasons. Census data sent the tip of the injury iceberg, as they do not describe the that yield accurate population denominators for calculating preponderant morbidity prevalence of the country. rates of injury in most cities in South Africa provide deci- The NIMSS statistics to a large extent reflect the gender sion and policy makers with quality information for mobi- trends in international injury statistics, indicating that in lizing resources toward the prioritization of injury preven- South Africa males are more likely than females to be the tion and safety promotion at both national and local levels victims of nonnatural deaths. Indeed, it would appear that of government. The experience of South Africa has indi- the gender divide in terms of injury prevalence is even more cated that although national data collection is imperative pronounced in South Africa than internationally. for the development of national injury profiling, a city focus Specifically, the NIMSS statistics indicate that, in the cities appears to expedite data collection, analyses, and delivery to participating in the NIMSS, 80 percent of all victims of non- decision makers aiming at preventive action. Injury preven- natural deaths are males (Donson and Van Niekerk 2002). tion practitioners have therefore reversed earlier attempts to These statistics also indicate that during 2001, males secure full coverage at the national level and refigured their accounted for 82.4 percent of all suicide cases. targets to data-driven prevention in cities. Ultimately, injury In the South African cities of Pretoria, Durban, Cape prevention researchers and practitioners believe that illus- Town, and East London, where injury-mortality surveillance trating the utility of injury surveillance at the city level will systems provide full coverage, injury rates indicate the dire inform the prioritization of a national injury prevention magnitude of injury as a disease burden. Pretoria, the capi- agenda. tal city of South Africa and home to a population of In Uganda, as in South Africa, violence and unintentional 2,043,500 (Sukhai and Matzopoulos 2002) had an all-injury injury are a major cause of disease burden and constitute the death crude rate of 136 per 100,000 in 2001. Although this fifth leading cause of premature death nationally (Uganda, figure pales in comparison to injury death estimates in Ministry of Health 2000). The magnitude of these burdens Kampala (the capital city of Uganda), the rate attests to the was measured in a survey conducted in Kampala by the prevalence of injury resulting in death in one of South Ministry of Health. In one section of Kampala 119 fatal Africa's cities with the most resources. Such figures continue injuries were reported for a population of 10,982 for the to call attention to the burden of injury in other South five-year period 1992 to 1996, resulting in a mortality rate of African cities, where crude injury mortality rates are equally 220 per 100,000 people per year. In the preceding six disturbing. In the west coast city of Cape Town, the tourist months, 138 injuries leading to disabilities were reported, hotspot of South Africa, the all-injury mortality rate of 170 giving an incidence of injuries leading to disabilities of Violence and Injuries | 365 23 per 1,000 people per year. The incidence of nonfatal also constituted 51 percent of all injuries from which this injuries was 114 per 1,000 per year. Furthermore, 312 dis- age group fully recovered, pointing to the importance of abilities as a direct result of injury were recorded by the burns in the population (see tables 24.1 to 24.5). survey. This translates into an injury-caused disability rate In the rural district of Mukono, in the southern part of of 2.8 per 1,000 per year. The overall incidence of all the the country, 34 fatal injuries were reported to have occurred injuries was 116 per 1,000 per year. Road traffic crashes con- in a population of 7,427 people during the five-year period stituted the primary cause of injury deaths and disabilities. preceding the ministry's survey. These figures translated into Interpersonal violence involving the use of firearms was the an average annual injury mortality rate of 92 per 100,000 second leading cause of injury mortality in this urban persons (see table 24.1). The prevalence rate of disabilities community. Burns were the leading cause of severe (fatal or due to injury was 0.7 percent. Drowning was the leading disabling) injuries in children age 10 years or younger. Burns cause of injury mortality in this rural district, followed by Table 24.1 Cause of Injury by Outcome: Mukono District, Uganda Outcome of injury Fatal injuries, 1994­98 Prevailing disabilities Fully recovered injuries (n 34) due to injury (n 55) past 6 months (n 575) Cause (no.) (%) (no.) (%) (no.) (%) Traffic 6 18.0 19 35.0 72 13.0 Burns 2 6.0 6 11.0 71 12.0 Stabs/cuts 4 6.0 9 16.0 197 34.0 Blunt force 5 15.0 6 11.0 109 19.0 Poison 1 3.0 0 0.0 2 0.3 Drowning 9 27.0 0 0.0 3 0.5 Animal bites 2 6.0 0 0.0 29 5.0 Falls 1 3.0 13 24.0 73 13.0 Gunshots 2 6.0 0 0.0 0 0.0 Other 2 6.0 2 4.0 19 3.0 Source: Kobusingye, Guwatudde, and Lett 2001. Table 24.2 Cause of Injury by Outcome: Kawempe Division, Kampala District, Uganda Outcome of injury Fatal injuries, 1992­96 Prevailing disabilities due Fully recovered injuries (n 119) to injury (n 312) past 6 months (n 478) Cause (no.) (%) (no.) (%) (no.) (%) Traffic 55 46.0 122 39.0 139 29.0 Burns 11 9.0 36 12.0 113 24.0 Stabs/cuts 1 0.8 21 7.0 73 15.0 Blunt force 4 3.0 15 5.0 30 6.0 Poison 9 8.0 7 2.0 4 0.8 Animal bites 6 5.0 18 6.0 14 3.0 Falls 4 3.0 39 13.0 85 18.0 Gunshots 24 20.0 35 11.0 7 1.0 Other 5 4.0 19 6.0 13 3.0 Source: Kobusingye, Guwatudde, and Lett 2001. 366 | Brett Bowman, Mohamed Seedat, Norman Duncan, and Olive Kobusingye Table 24.3 Cause of Injury by Outcome: Gulu District, Uganda Death rate per 10,000 Disability rate per 10,000 Injury rate per 10,000 Cause of injury (no.) (rate) (no.) (rate) (no.) (rate) Gunshots 168 32.8 87 17.0 288 56.2 Stabs/cuts 71 13.9 59 11.5 183 35.7 Blunt force 41 8.0 112 21.9 201 39.2 Land mines 26 5.1 36 7.0 70 13.7 Poisoning 22 4.3 24 4.7 58 11.3 Dog, snake, etc., bites 9 1.8 16 3.1 63 12.3 Traffic 8 1.6 55 10.7 88 17.2 Burns 7 1.4 19 3.7 63 12.3 Drowning 5 1.0 0 0.0 5 1.0 Falls 1 0.2 76 14.8 117 22.8 Other 39 7.6 95 18.5 171 33.4 All causes 397 77.5 579 113.0 1,307 255.1 Source: Kobusingye, Guwatudde, and Lett 2001. Table 24.4 Top Three Causes of Injury, by Age, Rural District (Mukono), Uganda Severe injuries Recovered injuries Age group n Causes No. % n Causes No. % 15 Falls 5 33 182 Burns 45 25 10 Traffic 4 27 Cuts/stabs 42 23 Burns 2 13 Falls 36 20 17 Burns 5 29 126 Cuts/stabs 36 29 10­19 Traffic 3 18 Blunt force 24 19 Drowning 3 18 Traffic 17 13 17 Traffic 5 29 103 Cuts/stabs 49 48 20­29 Cuts/stabs 3 18 Blunt force 21 20 Drowning 3 18 Traffic 16 16 11 Traffic 5 45 84 Cuts/stabs 34 40 30­39 Falls 2 18 Traffic 19 23 Drowning 2 18 Blunt force 17 20 7 Traffic 3 43 40 Cuts/stabs 18 45 40­49 Cuts/stabs 2 29 Blunt force 6 15 Burns 1 14 Traffic 5 13 15 Traffic 5 33 37 Cuts/stabs 17 46 50 Blunt force 4 27 Falls 6 16 Cuts/stabs 3 20 Traffic 5 14 n.a. n.a. n.a. n.a. 3 Falls 1 33 Not stated Cuts/stabs 1 33 Blunt force 1 33 82 Traffic 25 30 575 Cuts/stabs 196 34 Total Falls 13 16 Blunt force 108 19 Cuts/stabs 11 13 Falls 72 13 Source: Kobusingye, Guwatudde, and Lett 2001. Note: n.a. not applicable. Violence and Injuries | 367 Table 24.5 Top Three Causes of Injury, by Age, Urban Division (Kawempe), Uganda Severe injuries Recovered injuries Age group n Causes No. % n Causes No. % 58 Burns 24 41 168 Burns 85 51 10 Traffic 14 24 Falls 31 19 Falls 7 12 Cuts/stabs 24 14 71 Traffic 25 35 90 Traffic 28 31 10­19 Falls 1 15 Falls 22 24 Burns 8 11 Cuts/stabs 20 22 126 Traffic 56 44 105 Traffic 48 46 20­29 Gunshots 27 21 Cuts/stabs 14 13 Animal bites 10 8 Blunt force 13 12 96 Traffic 50 52 61 Traffic 27 44 30­39 Gunshots 20 21 Falls 12 20 Falls 5 5 Cuts/stabs 6 10 40 Traffic 19 48 25 Traffic 11 44 40­49 Gunshots 6 15 Falls 4 16 Falls 4 10 Burns 3 12 50 33 Traffic 11 33 13 Traffic 4 39 Falls 9 27 Falls 4 39 Cuts/stabs 3 9 Cuts/stabs 3 23 Not stated 7 Traffic 2 29 16 Cuts/stabs 4 25 Burns 1 14 Traffic 3 19 Blunt force 1 14 Falls 2 13 Total 431 Traffic 117 27 478 Traffic 139 29 Gunshots 59 14 Burns 113 24 Burns 47 11 Cuts/stabs 73 15 Source: Kobusingye, Guwatudde, and Lett 2001. road traffic accidents. Falls were the most common cause of These figures serve well as a picture of the alarmingly severe injuries in children below 10 years of age (33 percent), high prevalence and magnitude of death due to injury in followed by traffic (27 percent) and burns (13 percent). Uganda and South Africa. Furthermore, these statistics In Gulu, a rural district in northern Uganda that is expe- describe and underscore the importance of data as the basis riencing a protracted armed rebellion, 1,475 households for prioritizing interventions aimed at the prevention of with 8,595 people were surveyed. Seventy-three percent of violence and injury in Africa. The establishment of similar the population lived in temporary housing, 46 percent were injury surveillance systems in other parts of the continent internally displaced persons living in camps, and 81.3 per- would thus contribute to a more complete profile of conti- cent of the total population was under 35 years. These con- nental injury patterns. This more comprehensive picture ditions are being experienced in several countries in Africa, would allow for inclusively prioritizing prevention pro- and the pattern of injuries has not been well described. grams for those countries not yet identified as primary Fourteen percent of the population were injured annually, injury prevention targets in Africa. gunshots being the leading cause of injury mortality. The annual injury mortality rate was 7.7 per 1,000 (95 percent confidence interval [CI], 7.01­8.49), and the annual injury RISK FACTORS disability rate was 11.3 per 1,000 (95 percent CI, 10.4­12.2). Only 4.5 percent of the injured were reported to be combat- Throughout this review of the magnitude of injury mortal- ants. Fifty percent of the injured received first aid, whereas ity, the data have implied various risk factors. In response to only 13 percent of those who died reached the hospital. The the range of injury risk factors identified within myriad injury fatality rate is three and a half times higher than that studies, the WHO has developed a factor matrix that in urban Uganda, and more than eight times the global aver- captures its typology of injury risk. These risk factors, which age (Lett, Kobusingye, and Ekwaru 2006). the WHO, in its ecological typology for grouping risk 368 | Brett Bowman, Mohamed Seedat, Norman Duncan, and Olive Kobusingye factors, refers to as "individual" factors, do not remain dis- workplace, places of worship, and schools. These factors creet through the complex causal paths that result in vio- include community integration or cohesion, the availability lence. So, for example, being a male is associated with being of firearms, the presence of gangs, neighborhood density, a risk factor for being both the perpetrator and the victim of policing, and drug trafficking. a violent (and in the case of the victim) fatal act. These risk factors are enmeshed within the complex relational Social Cohesion. The degree of community cohesion and exchanges in the production of fatal injury. They interact functionality is strongly related to the rates of various forms both within and between the broader matrices of the risk of violence and injury. Studies have found a cyclical relation factors of fatal injury. Following the ecological typology between the levels of violence and community dysfunction proposed by the WHO (2002a), relationships, community, and lack of cohesion in Jamaica (WHO 2002a). Along simi- and society form another set of risk factors. lar lines, other studies have found that communities with the lowest levels of intimate partner violence were those characterized by levels of cohesion and functionality that Relationship Risk Factors allowed for sanctions against intimate partner violence and Relationship factors can be defined as those factors related effective victim support (WHO 2002a). In South Africa, to close personal relationships, such as family, peer, and inti- Keikelame and Ferreira (2000) also found a strong relation- mate partner relationships, that increase the likelihood of ship between community dysfunction and elder abuse. violence and injury. Drug Trafficking. Internationally, violence has been found Family Relationships. Family relationships appear to be to be closely associated with drug trafficking. For example, central to the development and occurrence of a broad range in Brazil, research has shown that drug trafficking is respon- of injuries and acts of violence. For example, violent behav- sible for a significant proportion of homicides and injuries ior during adolescence has been linked to conflictual (WHO 2002a). In other parts of South and Central parental relationships (WHO 2002a). Furthermore, research America, studies have also found that adolescents involved has found that the physical abuse and neglect of children in drug dealing exhibit higher levels of violence than their were strong predictors of the latter's later arrest for violence. peers who are not (WHO 2002a). Similarly, other studies have found that the harsh physical punishment of children by their parents at the age of eight Societal Factors years appeared to increase the former's chances of arrest for violence up to the age of 30 years (WHO 2002a). Moreover, This set of risk factors are those broad social factors "that it also appeared to increase the likelihood that these children create an acceptable climate for violence, those that reduce would become physically abusive parents themselves. inhibitions against violence, and those that create and sustain gaps [and tensions] between different groups or Peer Relationships. Peer relationships, particularly during countries" (WHO 2002a, 13). These include health, educa- adolescence, appear to be strongly related to violent behav- tional, economic, and political factors and social policies ior and intentional injuries. For example, researchers have and practices that result in high levels of social, political, and found a strong relationship between violence and having economic inequality. Prevailing cultural values and practices delinquent friends (WHO 2002a). However, the causal also form part of this set of societal risk factors. direction of this relationship, that is, whether having delin- quent friends leads to violence or whether violence results in Economic Inequalities. The WHO (2002a) argues that the establishment of friendships with delinquent friends is there is a manifest relationship between poverty and levels not clear. of violence and injury. This position is borne out by a range of studies. For example, various studies conducted in Africa report a link between poverty and the risk of child- Community Factors hood burns (Mbembati, Maseru, and Leshabari 2002). This repertoire of risk factors includes those related to Additionally, as reported earlier, 87.9 percent of all road traf- various community contexts that may increase the likeli- fic deaths occur in LMICs (WHO 2002a). Then too, the hood of violence and injury, such as the neighborhood, the WHO (2002a) indicates that although the composite rate of Violence and Injuries | 369 homicide, suicide, and war-related deaths for high-income Prevention of Road Traffic Injuries countries is 14.4 per 100,000, the rate for LMICs is 32.1 per In the prevention of RTIs the focus could be on proven 100,000, more than twice as high as that for HICs. In this measures, such as encouraging pedestrians to walk facing regard, studies have found that the gross domestic product traffic, introducing bright and visible clothing, enacting across countries is negatively related to violence (WHO speed limits, persuading drivers to use headlights during the 2002a). In South Africa, Budlender (2000) has also reported daytime, and promoting the use of approved disc brakes. a link between poverty and death and injury due to crime. Environment-based measures for the prevention of RTIs However, in many cases, this relationship is substantially include vigilant and consistent policy and enforcement moderated by income inequalities. In effect, significant measures; the introduction of red light cameras; the use of income inequalities are associated with high rates of vio- traffic circles; the bright illumination of crossings; and the lence, even in high-income contexts. To a certain extent, this provision of adequate sidewalks for pedestrians, who are finding could perhaps account for the high levels of violence among the most vulnerable road users. Other prevention in South Africa, given its substantial income disparities measures include segregating public vehicles from private (May, Woolard, and Klasen 2000). transport vehicles; providing adequate separate lanes for heavy vehicles, cyclists, and pedestrians; and introducing a Culture. To a large extent, people's responses to their envi- flexible time arrangement for car-free zones for pedestrians ronment are influenced by culture, which is reflected in the (Mohan 2003). values and norms of a society. For this reason, culture is also Control measures aimed at reducing the severity of RTIs considered to play an important role in the levels of violence may include requiring the use of protective clothing and in a particular society. Research shows that cultures that helmets for cyclists, the use of seatbelts, and the develop- support values that endorse violence and that do not pro- ment of safer dashboards in motor vehicles. Key measures vide nonviolent alternatives to resolve conflicts appear to for the effective management of traffic-related injuries have higher rates of violence. More specifically, a number of include rapid responses to crashes in the form of first aid for studies have found that the risk of violence against women the injured, elimination of fuel leakage, and clearance of is increased in cultures that endorse male dominance and accident scenes, in addition to effective policing systems and aggression (WHO 2002a). Additionally, the WHO (2002a) accessible hospital and rehabilitation centers for the injured argues that contemporary cultural norms that valorize (Forjuoh 1996). youthfulness is a significant contributory factor in the elevated levels of elder abuse. Violence Prevention Violence prevention measures may be universal, selected, or STRATEGIES FOR CONTROL: WHAT indicated in nature. Whereas universal interventions PRACTICES EXIST? are aimed at the total population without consideration of individual risks (for example, community-wide media The preceding analysis and the work of international agen- campaigns), selected interventions target those at risk for cies, including the WHO, point to the importance of estab- violence (for example, training single parents in parenting lishing injury surveillance systems that can provide timely, skills and anger management). Other interventions should accurate, and quality data for informing prevention policies be directed at those who have already engaged in violent and practices and serve as the backbone of national preven- behavior (for example, rehabilitation programs for perpe- tion programs. Without surveillance as a method and a trators of domestic violence). process, health care planners are unable to prioritize the Selected interventions that are keenly attuned to risk allocation of resources. Thus, a national program may focus factors that include biological factors, psychological and on strengthening capacities for data collection and preven- behavioral characteristics, family influences, peer influences, tion methodologies, encouraging research on the determi- income, gender, age, geographical location, and various nants and costs of injuries and their prevention, encourag- other context-specific variables have been shown to be effec- ing intersectoral and cross-disciplinary collaboration, and tive in preventing violence (Berger and Mohan 1996). enhancing responses for the prevention, control, and man- Following the ecological approach of the WHO (2002a), agement of injuries. violence prevention targeting the vulnerable 12-to-19-year 370 | Brett Bowman, Mohamed Seedat, Norman Duncan, and Olive Kobusingye age cohort may, for instance, include social development the stimulation and encouragement of safer working tech- programs, academic enrichment programs, mentoring ini- niques in the workplace (WHO 2002b). tiatives, extracurricular activities, and attempts to reduce Measures for the prevention of drowning as recom- poverty and income inequality (WHO 2002a). The causal mended by the WHO (2002b) include the development of pathways of violence are complex, and the prevention of strategies to ensure safe water transport and commutation, interpersonal violence should be multisectoral and multidi- the sensitization of policy makers to the benefits of life jack- mensional and therefore include all the identified risk ets and other flotation devices to the protection of children factors that are specific to the context in which such violence and other vulnerable users near water, and the provision of occurs (Zwi et al. 1996). A survey study conducted by adequate life protection services (in the form of adult super- Butchart, Kruger, and Lekoba (2000), in which community vision or qualified life guards) for recreational swimming perceptions of violence yielded qualitative information, times and places (WHO 2002b). provides an alternative means of gaining rich information on the causes and possible prevention of violence. This information lends itself to analysis that may uncover the CONCLUSION nuances and context specificities often overlooked by more broadly based epidemiological studies. Violence and injury are salient and definitive threats to the health of African nations, and the globe. Consequently, the inclusion of violence and injury prevention curricula across Burns, Falls, and Drowning Prevention the full spectrum of primary, secondary, and tertiary educa- The WHO (2002b) has provided the most comprehensive tion platforms, as well as the training of African injury set of recommendations and guidelines for the prevention prevention specialists, should be seen as vitally important. of burns, falls, and drowning to date. Although most of Fortunately, it appears as though these areas are receiving these prevention suggestions are formulated in relation to growing attention in some African countries. For instance, the Southeast Asia region, their generic applicability to other in a positive development, the WHO Injuries and Violence LMICs cannot be underestimated. Country-specific factors Prevention Department collaborates with Mozambican must, however, be included in prevention strategies that counterparts to enhance the safety and security of take local risk factors and available resources into account. Mozambique's 20 million people. The Mozambique Project, Country-specific studies should tailor these guidelines to as a national action plan, includes the development of their local contexts; for example, studies in South Africa national violence prevention policies, the detailed country- suggest that the unsafe use of paraffin is significantly associ- wide analysis of injuries, the development of a network of ated with burn injuries (Steenkamp, van der Merwe, and prevention practitioners through training and awareness de Lange 2002). Paraffin management strategies must there- campaigns, and an initiative aimed at controlling the supply fore inform burn prevention strategies for the country. and distribution of small arms. Perhaps the most powerful Recommendations for the prevention of burn injuries illustration of the recent prioritization of injury prevention include the provision of stable lamps and stoves in LMIC and safety promotion for combating the devastating social contexts, effective training of personnel in evacuation and economic effects of injuries and violence on the African processes following fires in the workplace, the installation of continent is embodied by a resolution passed by the heads of fire and smoke alarms in public buildings, the regulation or states of the African Union (signed in Maputo in 2003). The elimination of dangerous fireworks, the greater use of resolution declared 2005 the African Year of Violence flame-resistant fabrics, and the promotion of the use of cold Prevention and endorsed the nine recommendations of the water in the treatment of burns at the tertiary prevention World Report on Violence and Health (WHO 2002a), which level (WHO 2002b). urged states to implement multisectoral national plans Falls could be best prevented through the use of accom- of action to limit injury and violence and to enhance data modating and soft materials, such as mud and sand, in the collection systems. design of playgrounds; the provision and implementation of In South Africa, the presidential Lead Programme on safety regulations for places in which children most fre- Crime, Violence and Injury is focused on the development quently play; legislating for safer designs for fall-vulnerable of injury surveillance methods; good practices for the structures, such as railings and grab bars on balconies; and prevention, control, and management of injuries; the Violence and Injuries | 371 implementation of injury prevention strategies through Forjuoh, S. N., A. B. Zwi, and C. J. Romer. 1996. Injury Control in Africa: Proceedings of Round Table Session and Associated Meetings Held at the training; the strengthening of current bottom-up and top- Third International Conference on Injury Prevention and Control, down initiatives; and the stimulation of prevention policies. Melbourne, Australia, February 18­22, 1996. Pittsburgh: University of A major focus of the program is to chart the contextual, Pittsburgh. social, process, and content factors that influence data uptake Holder, Y., M. Peden, E. Krug, J. Lund, J. Gururaj, and O. Kobusingye, eds. 2001. Injury Surveillance Guidelines. Geneva: WHO. and the stimulation of prevention policies and practices Jansson, B., and L. Svanstrom. 1999. "Injury as a Public Health Problem: (Seedat 2002). Looking Behind the Figures." In Safety Promotion Research, ed. Injury prevention may thrive while a national program L. Laflamme, L. Svanstrom, and L. Schelp, 43­62. Kristianstads: Kristianstads Boktryckeri AB. focuses on developing systems to record the magnitude and Keikelame, J., and M. Ferreira. 2000. "Mpathekombi, ya Bantu abadala: risks of injuries, evaluating and documenting promising Elder Abuse in Black Townships on the Cape Flats." Human Sciences and effective practices, stimulating the wide-scale imple- Research Council and University of Cape Town Centre for Gerontology, Cape Town. mentation of these practices, and encouraging the develop- Kobusingye, O., D. Guwatudde, and R. Lett. 2001. "Injury Patterns in Rural ment of evidence-led policies and associated funding and Urban Uganda." Injury Prevention 7: 46­50. decisions. To ensure the long-term development of the Lerel, L. 1994. "The Epidemiology of Fatal Childhood Burns." South injury and violence prevention sector in Africa, however, African Medical Journal 84 (3): 169­70. intersectoral and multidisciplinary collaboration, technical Lett, R. R., O. C. Kobusingye, and P. Ekwaru. 2006. "Burden of Injury dur- ing the Complex Political Emergency in Northern Uganda." Canadian cooperation, and partnerships between government and Journal of Surgery 49 (1): 51­57. civil society are vital. Marysse, S. 2003. "Regress and War: The Case of the DR Congo." European Journal of Development Research 15 (1): 73­98. Mathers, C. D., R. Sadana, J. A. Salomon, C. J. L. Murray, and A. D. Lopez. 2001. "Healthy Life Expectancy in 191 Countries, 1999." Lancet 357: REFERENCES 1685­91. Matzopoulos, R., ed. 2002. A Profile of Fatal Injuries in South Africa: Third Berger, L., and D. Mohan. 1996. Injury Control: A Global View. New York: Annual Report of the National Injury Mortality Surveillance System. Oxford University Press. Cape Town: MRC Press. Bowley, D. M. G., N. K. Parmar, and K. D. Boffard. 2004. Burdens of May, J., I. Woolard, and S. Klasen. 2000. "The Nature and Measurement Disease in Southern Africa. Lancet 363: 1508­9. of Poverty and Inequality." In Poverty and Inequality in South Africa: Budlender, D. 2000. "Human Development." In Poverty and Inequality in Meeting the Challenge, ed. J. May. Cape Town: David Phillip. South Africa: Meeting the Challenge. Cape Town: David Philip. Mbembati, N. A., L. M. Maseru, and M. T. Leshabari. 2002. "Childhood Burrows, S., B. Bowman, R. Matzopoulos, and A. van Niekerk, eds. 2001. Burn Injuries in Dar es Salaam: Patterns and Perceptions of A Profile of Fatal Injuries in South Africa 2000: Second Annual Report of Prevention." African Safety Promotion 1 (1): 42­45. the National Injury Mortality Surveillance System. Cape Town: MRC Meel, B. 2003. "Determinants of Suicide in the Transkei Sub-Region of Press. South Africa." Journal of Clinical Forensic Medicine 10: 71­76. Burrows, S., M. Vaez, A. Butchart, and L. Laflamme. 2003. "The Share of Mock, C. N., S. N. Forjuoh, and F. P. Rivara. 1999. "Epidemiology of Suicide in Injury Deaths in the South African Context: Transport-Related Injuries in Ghana." Accident Analysis and Prevention Sociodemographic Distribution." Public Health 117 (1): 3­10. 31: 359­70. Butchart, A., J. Kruger, and R. Lekoba. 2000. "Perceptions of Injury Causes Mohan, D. 2003. "Typical Crash Analysis." Paper presented at the Fifth and Solutions in a Johannesburg Township: Implications for Injury Control and Traffic Safety Training Course, Johannesburg, July Prevention." Social Science and Medicine 50: 331­44. 21­25. Cliff, J., and A. R. Noormohamed. 1993. "The Impact of War on Moshiro, C., R. Mswia, K. Alberti, D. Whiting, N. Unwin, and P. Setel. 2000. Children's Health in Mozambique." Social Science and Medicine 36 (7): "The Importance of Injury as a Cause of Death in Sub-Saharan Africa: 843­48. Results of a Community-Based Study in Tanzania." Public Health 115: De Wet, B. 1993. "Reflect to Be Seen." Trauma Review 1 (3): 2. 96­102. Donson, H., and A. Van Niekerk. 2002. "Suicide." In A Profile of Fatal National Center for Injury Prevention and Control (NCIPC). 2003. Injuries in South Africa: Third Annual Report of the National Injury "Intimate Partner Violence Fact Sheet." Atlanta: NCIPC. http://www. Mortality Surveillance System, ed. R. Matzopoulos, 31­34. Cape Town: cdc.gov/ncipc/factsheets/ipvfacts.htm, accessed September 12, 2003. MRC Press. Nattrass, N., and J. Seekings. 2001. "Two Nations? Race and Economic Elliot, G. 2000. "Mozambique: Development through De-Mining." South Inequality in South Africa Today." Daedalus 130: 45­70. African Journal of International Affairs 7 (1): 97­105. Odero, W., M. Khayesi, and P. M. Heda. 2003. "Road Traffic Injuries in Elliot, G., and G. Harris. 2001. "A Cost-Benefit Analysis of Landmine Kenya: Magnitude, Causes and Status of Intervention." Injury Control Clearance in Mozambique."Development Southern Africa 18 (5): 625­33. and Safety Promotion 10 (1­2): 53­61. Forjuoh, S. N. 1995. "Pattern of Intentional Burns to Children in Ghana." Peden, M., and K. McGee. 2003. "The Epidemiology of Drowning Child Abuse and Neglect 9 (7): 837­41. Worldwide." Injury Control and Safety Promotion 10 (4): 195­99. ------. 1996. "Burn Repetitions in Ghanaian Children: Prevalence, Peden, M., K. McGee, and G. Sharma. 2002. The Injury Chart Book: Epidemiological Characteristics and Socioenvironmental Factors." A Graphical Overview of the Global Burden of Injuries. Geneva: Burns 22 (7): 539­42. WHO. 372 | Brett Bowman, Mohamed Seedat, Norman Duncan, and Olive Kobusingye Peltzer, K., and W. Renner. 2003. "Superstition, Risk-Taking, and Risk Third Annual Report of the National Injury Mortality Surveillance Perception of Accidents among South African Taxi Drivers." Accident System, ed. R. Matzopoulos, 31­34. Cape Town: MRC Press. Analysis and Prevention 35: 619­23. Uganda, Ministry of Health. 2000. Health Management Information System Schlebusch, L., and B. A. Bosch, eds. 2000. "Suicidal Behaviour." 4. Report. Kampala: Ministry of Health. Proceedings of the Fourth South African Conference on Suicidology. Venter, P. 1998. "The Young Pedestrian in Traffic." Trauma Review 6 (3): Durban: University of Natal. 8­9. Seedat, M. 2002. "Extending the Boundaries of Injury Prevention Theory, WHO (World Health Organization). 2002a. World Report on Violence and Research and Practice in Africa." African Safety Promotion 1 (1): 5­15. Health. Geneva: WHO. Smith, G. S., and P. Barss. 1991. "Unintentional Injuries in Developing ------. 2002b. Injuries in South-East Asia Region: Priorities for Policy and Countries: The Epidemiology of a Neglected Problem." Epidemiology Action. Geneva: WHO. Reviews 13: 228­66. Zwi, A., S. Murugusampillay, B. Msika, et al. 1993. "Injury Surveillance in Steenkamp, W. C., A. E. van Der Merwe, and R. de Lange. 2002. "Burn Zimbabwe: A Situation Analysis." Ministry of Health and Child Injuries Caused by Paraffin Stoves." South African Medical Journal 92 Welfare, Zimbabwe, and London School of Hygiene and Tropical (6): 445­46. Medicine, London. Stohl, R. J. 2002. "Under the Gun: Children and Small Arms." African Zwi, A., S. Forjuoh, S. Murugusampillay, W. Odero, and C. Watts. 1996. Security Review 11 (3): 17­25. "Injuries in Developing Countries: Policy Response Needed Now." Sukhai, A., and R. Matzopoulos. 2002. "Appendix 1: Inter-city and Transactions of the Royal Society of Tropical Medicine and Hygiene 90: Regional Comparisons." In A Profile of Fatal Injuries in South Africa: 593­95. Violence and Injuries | 373 Index Boxes, figures, and tables are indicated by b, f, and t, respectively. abortion mental disorders and Benin government policy on, 231­232, 232f depression, 334 adult mortality rates in, 38 postabortion care, 229 primary prevention of, 341­342 life expectancy projections, after AIDS, unsafe, 229 prevalence of, 318t 65­66, 65f ACT (artemisinin combination Alcohol Use Disorders Identification Test population projections after AIDS, therapy), 208 (AUDIT), 330 62, 62f acute respiratory infections (ARI), Alma Ata Declaration on Primary Health safe water and sanitation in, 98 149­151 Care (1978), 228 women's life expectancy after AIDS, 67f adolescents and young adults alternative medicine. See traditional Bill & Melinda Gates Foundation, 7 causes of death, 45, 57 medicine bipolar disorders, 332 vital records on, 46, 47t Alzheimer's disease, 354 birth. See childbirth cholesterol levels in, 258 anemia blackflies and onchocerciasis, 215 physical activity of, 252 developmental delays and, 138 vector control, 217­218, 218f, 220t smoking and, 248, 249t economic effects of, 88, 95 Black Risk Factors Study (BRISK) on violent behavior and injuries of, 369 malaria and, 196, 203, 205, 206 hypertension among blacks in Adult Morbidity and Mortality Project in women, 95, 229 South Africa, 316 (AMMP), 33, 45, 316 antidepressants, 342 blindness, 131, 135, 137 adverse drug reactions to commonly used antiretroviral drugs (ARV) river blindness. See onchocerciasis antimalarial drugs, 203 HIV/AIDS and, 238, 242, 243­244, 295 blood lipids. See dyslipidemia aerobic exercise, 251­252 neurological complications, 358 blood pressure. See cardiovascular disease aflatoxin, 296, 297 TB patients with HIV, 187, 189 (CVD), subheading: hypertension African Programme for Onchocerciasis "3 by 5" Initiative of WHO for, 322 blood transfusions Control (APOC), 216, 217, 219, 220, APOC. See African Programme for HIV/AIDS and, 3, 241­242 220t, 221 Onchocerciasis Control malaria and, 203 African Year of Violence ARI. See acute respiratory infections Board of Science and Technology for Prevention (2005), 371 artemisinin combination therapy International Development aging. See elderly persons (ACT), 208 (BOSTID) study of lower agricultural practices ARV. See antiretroviral drugs respiratory tract infections, malaria and, 199 asthma. See respiratory diseases of children 150, 153 productivity and nutrition attention-deficit/hyperactivity BOMA. See Burden of Malaria in Africa improvements, 103 disorder, 343 (BOMA) project AIDS. See HIV/AIDS AZT, 243, 295 Botswana alcohol consumption birth rate projections after AIDS, 63, 64f burden of, 3 bacillus Calmette-Guérin (BCG), 170, 189 HIV/AIDS in, 239 consequences of, 339­341 bacteria prevalence estimates and demographic CVD and, 317 as cause of meningitis, 172 impact, 71­72, 72f data specific to Sub-Saharan Africa, 335 as leading cause of lower respiratory tract women's life expectancy and, 66 developmental disabilities and, 139 infections, 152, 153 malaria in, 201 esophageal cancer and, 299 bacterial meningitis, 136­137, 353 vital records in, 45 hypertension and, 255 BCG (bacillus Calmette-Guérin), 170, 189 breast cancer, 293­294, 300 375 breastfeeding coronary heart disease, 310­311, 314, GBD process for estimating, 44, 47­48, cancer and, 294 315, 315t 48f, 52 diabetes and, 272 costs of interventions, 322, 323 leading causes of death transmission of HIV and, 242 crude mortality vs. age-standardized by age, 54­55, 55­56t undernutrition of children and, 92, 93, mortality, 307, 308f regional comparisons, 54, 54t 100, 101f data sources on, 307 by sex, 53­54, 53t, 55­56, 56t burden of disease. See specific diseases diabetes and, 317 life tables, construction of, 48­49, 49f, 61 and conditions diet contributing to, 253 verbal autopsies and. See verbal autopsies Burden of Malaria in Africa (BOMA) See also obesity vital records showing, 45­47, 46­47t project, 200, 208 dilated cardiomyopathy, 312 census data. See U.S. Census Bureau Burkina Faso drug treatment of, 321­322 Center for International Earth Science adult mortality rates in, 38 dyslipidemia and, 258­260, 259t Information Network (CIESIN) on alcohol consumption in, 335 dysrhythmias, 313 distribution of child HIV/AIDS in, 71 endomyocardial fibrosis, 312 undernutrition, 93­94, 94f malaria in, 206 epidemiological transition in, 306­307, cerebral malaria, 136, 200, 203, 354 onchocerciasis control program in, 217 306t, 316 cerebral palsy, 140­141, 353 Burkitt's lymphoma, 298 epidemiology of, 309­314 See also neurological disorders burn-related injuries, 364, 371 genetic determinants of, 314, 319 cerebrovascular disease, 279, 309­310, 354 Burundi HIV-related, 4, 312­313 See also stroke Household and Living Standards hyperlipidemia, 317 cervical cancer, 292­293 Survey, 330 hypertension, 254­257, 256t, 315­317 Chad malaria in, 202 cost-effectiveness of interventions for, adult mortality rates in, 38 stomach cancer in, 295 257, 323 malaria in, 206 undernutrition of children in, 101 interventions for, 256, 257, 320, refugee populations of, 9 322, 323 child abuse, 338, 369 Cameroon prevalence of, 315­316, 316t, 355 childbirth adult mortality rates in, 37 stroke and, 355 See also maternal conditions cardiovascular disease in, 309 urbanization and, 255, 315­316, 316t access to skilled attendance, 231, diabetes in, 272, 276, 277, 278, 280 interventions, 319­320 231f, 233 HIV/AIDS in, 240 pericarditis, 314 emergency obstetric care, 233­234 hypertension in, 315 physical activity and, 317 low birthweight. See low birthweight life expectancy in, 277 prevalence and incidence of, 308­309 maternal death. See maternal mortality noncommunicable disease burden in, 252 prevention, 319­320 neonatal death. See child mortality rates physical activity in, 277 rheumatic heart disease (RHD), 311 preterm birth and malaria, 204­205 urban lifestyle and lack of physical risk factors for, 314­315t, 314­319 projected rates of, after AIDS, 63 activity in, 252 low birthweight and, 248 childhood health cancer, 289­301 screening for, 320, 321 See also specific diseases and conditions See also specific types sickle-cell disease, 313 burns, 364 burden of, 2 smoking and, 317 cancer and, 298 data sources on, 290­291, 300­301 causes of death, 43­58, 290t developmental disabilities, 126­128 epidemiological estimates of deaths cause-specific estimating, 52­53, 53f See also developmental disabilities caused by, 52 data sources on, 43­53, 76 diabetes, 270, 272 HIV/AIDS and, 4, 299 epidemiological estimates of, 49­52, 317f diarrheal diseases. See diarrheal diseases incidence of, 289, 290t, 291­292, 292f cancer, 52 drownings, 364 male vs. female, 291­292, 291t CVD, 305, 307, 308 epilepsy. See epilepsy screening, 293 diarrheal diseases, 50, 50t malnutrition. See stunting; tobacco-related cancers, 299­300 epilepsy, 353 undernutrition of children women and, 292­294 HIV/AIDS, 51­52 mental disorders, 335 cardiovascular disease (CVD), 305­323 lower respiratory infections, 51 treatment of, 343 age of onset, 315 malaria, 50, 50t neurological disorders, 355, 358 alcohol consumption and, 317 maternal mortality, 51 perinatal conditions. See perinatal burden of, 2­3, 305 suicide, 363 conditions as cause of death, 56, 305 vaccine preventable diseases, 50­51 respiratory disease. See respiratory measurement of, 307 war and violence, 52, 362­363 diseases of children cerebrovascular accidents, 309­310 epidemiological literature, review of, stunting. See stunting clustering of multiple risks, effect of, 261 44­45, 44t TB and, 182 376 | Index child mortality rates, 15­30 climate determinants of malaria, 153, hypertension in, 315 under age 5 mortality, 11 197, 198f TB in, 190 annual rate of change in, 19­20, 20f cognitive function cotrimoxazole treatment, 190 causes of, 44, 44t, 46, 46t fetal alcohol syndrome and, 139 cultural beliefs. See religious and diarrheal disease and, 107, 108f, malnutrition's effect on, 139 cultural customs 109f, 117 communicable diseases. See infectious and CVD. See cardiovascular disease estimated levels, 18­19, 18f, 25­29, communicable diseases 26f, 27­29t community-based rehabilitation and DALYs. See disability-adjusted life years HIV/AIDS and, 159, 239 developmental disabilities, 142­143 data collection, 7­8 interventions to reduce, 22 community-directed treatment with See also vital events registration lower respiratory tract infections and, ivermectin (CDTI), 219­220, adult mortality rates, 32­33 149, 152, 159 220t, 221 on cancer, 290­291, 300­301 malaria and, 202 community violence, 369 on causes of death, 43­53, 76 malnutrition and, 95­96, 96f comparison of Sub-Saharan Africa with child mortality rates, 16­18 MDGs for, 15, 117 other regions and countries on CVD, 307 measurement of, 78­80, 80f, 80t developmental disabilities prevalence, on developmental disabilities, 128­129, trends in, 19­20, 19f, 22, 107, 129, 130t 129­130t 126­127 HIV/AIDS prevalence, 238­239, 239t on diabetes, 268­270, 269­271t causes of death, 44, 44t, 56­57 mortality rates, 12, 12t, 13, 13t, 14f on diarrheal diseases, 108­110, 117­119 leading causes of death, 54­55, 55t TB prevalence, 183, 184f on injuries, 362 vital records on, 46­47, 46­47t Comprehensive Africa Agriculture maternal mortality rates, 32, 223­226 data available on, 15­16, 16f Development Programme, 104 on mental disorders, 330­331 measurement of, 78­80, 80f, 80t condoms. See HIV/AIDS model life tables, use of, 76 sources and methods for, 16­18 conflicts. See wars and conflicts on mortality rates, 32­33 study data used, 17 congenital disorders, 132­134, 133t on neurological disorders, 352 infant and neonatal mortality, 11 heart disease, 313 population projections after AIDS, 59­61 comparison of Sub-Saharan Africa congenital rubella, 135 recommendations for improving, 41 with other regions, 12, 13f, 13t congestive heart failure. See cardiovascular on undernutrition of children, 92 malaria and, 204 disease (CVD) deafness. See hearing loss MDGs on, 234 Congo, Democratic Republic of death. See child mortality; mortality rates; measurement of, 78­80, 80f, 80t cancer in, 295, 296 specific diseases tetanus and, 168 cardiovascular disease in, 312, 314 dehydration. See diarrheal diseases trends in, 126­127 low birthweight and, 248 dementia, 354 malaria and, 202, 206 child mortality rates in, 15, 15f Democratic Republic of Congo. See Congo, methodology for estimation of trends, conflict in, 8 Democratic Republic of 17­18, 22­25 diabetes in, 273 Demographic and Health Surveys (DHS), application of, 25 hypertension in, 315, 316 12, 76 linear splines and "knots," 23­24, 23f immunization rates in, 9 diarrheal diseases, 108, 115 weighted least squares with linear infant mortality rates in, 13 maternal mortality, 228, 229 splines, 23 malaria in, 203, 206 undernutrition of children, 92 weights, 24 TB recurrence in HIV-infected demographic context, 5­6, 6f chloroquine, 208 individuals in, 188 See also population trends cholera, 115, 116t undernutrition of children in, 101 demographic surveillance systems (DSS) cholesterol levels, 258­260, 259t war in, 363 malaria monitoring, 200, 206, 208 CVD and, 260, 315, 320 consanguineous marriages and congenital mortality measurement, 75, 76, 85 lifestyle and, 251 abnormalities, 132, 134 depression. See mental disorders chronic diseases Core Welfare Indicators Questionnaire developmental disabilities, 2, 125­147 See also specific diseases (CWIQ), 330 accidents as cause of, 140 burden of, 247 coronary heart disease, 310­311, 314, bacterial meningitis as cause of, 136­137 interventions recommended, 261­262 315, 315t burden of disease in childhood, 126­128 risk factors for, 246­262 See also cardiovascular disease (CVD) measurement of, 127­128 See also risk factors for cost-effective interventions. See specific causes of, 135­137 noncommunicable diseases diseases and conditions cognitive disabilities, 130 cigarette smoking. See tobacco use Côte d'Ivoire community-based rehabilitation and, 142 clean drinking water. See safe water, adult mortality rates in, 38 cost-effectiveness of, 143 access to cancer in, 290 congenital disorders, 132­134 Index | 377 developmental disabilities (Continued ) complications of, 277­279 mortality rate and, 107, 108f, 117 congenital rubella as cause of, 135 costs of treatment and care, 279­280, oral rehydration therapy (ORT), 107 congenital syphilis as cause of, 135 281t, 282, 282f poverty and, 114 cost-effectiveness of community-based CVD and, 317 refugees and, 9 rehabilitation, 143 data sources on, 268­270, 269­271t risk factors for, 114­115 data needs, 143 environmental factors for, 272 rural vs. urban sites, 115, 116t defined, 125 epidemiology of, 270­271 sanitation improvements and, 114­115, determinants and risk factors for, estimates of type 2 diabetes, 273 115f, 115t 132­134, 133t family history of, 277 water and sanitary facilities and, Down syndrome, 134 foot problems and, 278­279 114­115 drug use, effect of, 140 genetic factors for, 272 diet. See nutrition environmental exposures to toxins, hypoglycemia, 278 diphtheria 139­140 immunological factors for, 272 See also Expanded Program on fetal alcohol syndrome, 139 increasing burden of, 2 Immunization (EPI); vaccine- hearing loss, 131, 131t interventions, 280­283 preventable diseases HIV children and, 3, 135­136 organization of, 282, 282b burden of, 166­167 injuries as cause of, 140 kidney disease and, 278 vaccine for, 163, 168­169 interventions, 140­143 management of, 282, 282b directly observed therapy short course intrauterine growth retardation mortality associated with, 279 (DOTS), 184, 185­190, 187f (IUGR), 134 neuropathy, 278­279 disabilities. See developmental disabilities iodine deficiency and, 138 obesity and, 277 disability-adjusted life years (DALYs), 127, iron deficiency and, 138 physical activity and, 277 179, 358 lead exposure, effect of, 140 prevalence and incidence of, 318t Disease and Mortality in Sub-Saharan learning disabilities, 132 type 1 diabetes, 271­272 Africa (DMSSA-1), 1­2, 2b malaria and. See cerebral malaria type 2 diabetes, 272­277 conditions not covered in, 2­3, 179 measles as cause of, 137 prevention, 283­284 on diarrheal diseases, 109 medical model of, 125­126 projections from 2003 to 2025, 273, 276t displaced persons. See refugee populations motor disabilities, 131 quality of treatment and care, 283 domestic violence, 338, 363, 369 neonatal screening for, 141 retinopathy, 279 DOTS. See directly observed therapy neural tube defects, 134 risk factors for, 273­277 short course perinatal and neonatal conditions, 134 self-management of, 284 Down syndrome, 134 poverty and, 126, 138­139 type 1 diabetes, 267­268, 270, driving injuries. See traffic-related injuries prevalence and incidence, 128­132, 353 271­272, 284 drowning, 364, 371 childhood disability surveys, 128 type 2 diabetes, 267­268, 268t, 270, drug resistance comparison with other developing 272­277 DOTS and. See directly observed therapy countries, 129, 130t urban vs. rural differences, 276­277 short course (DOTS) reliable data, 128­129, 129­130t Diabetes Associations, role of, 284 malaria and, 199, 208 prevention levels, 140­142, 141f Diabetes Atlas, 284 TB and, 185 public health interventions, development Diabetes Control and Complication drugs of, 132, 133f Trial, 283 See also specific drugs and diseases social model of, 125­126 diarrheal diseases, 107­123 for depression, 342 tetanus as cause of, 137 burden of, 108 dispensing, 7 undernutrition and, 137­139 children under age of five, 115, 116f, 117 donated drugs vision loss, 131 data sources on, 108­110, 117­119 insulin, 284 vitamin A deficiency and, 138 estimating burden of, 108 ivermectin, 219 WHO's International Classification of etiology, 109­110, 119t for epilepsy, 357 Functioning, Disability and Health, morbidity, 108, 117t for hypertension, 257 126, 126f, 143 mortality, 108­109, 109f, 118t insulin. See insulin WHO's International Classification of reviews and estimations, 110­114, 110t, prevention of mental disorders, 341­342 Impairments, Disabilities and 111f, 112­114t for schizophrenia, 342 Handicaps, 126, 126f epidemiological estimates of deaths drugs, illegal diabetes, 267­284 caused by, 50, 50t burden of use of, 3 age and ethnicity as factors for, 276, 276f gender differences and, 115 consequences of use of, 339­341 age of onset, 272, 280 HIV/AIDS and, 108 developmental disabilities caused by, 140 care, 283­284 interventions, 116­117 trafficking as cause of injuries, 369 cerebrovascular disease and, 279 malnutrition and, 96 DSS. See demographic surveillance systems 378 | Index dysentery, 116 fall-related injuries, 364, 371 Ghana See also diarrheal diseases family violence. See domestic violence cardiovascular disease in, 305, 308, dyslipidemia, 258­260, 259t FAO. See Food and Agriculture 311, 312 dysrhythmias, 313 Organization children with disabilities in, 140 fetal alcohol syndrome, 139 diabetes in, 273, 280 economic growth and consequences, 4­5, 4f financing hypertension in, 257, 315 gross national income by country, 4, 5t for reproductive health, 232 lower respiratory tract infections in, 149 undernutrition and, 87­88, 95, 100­101, for vaccines, 174­175 malaria in, 202 102, 102t, 105f food. See nutrition nutrition in, 255 economic loss Food and Agriculture Organization (FAO) tobacco use in, 248 See also disability-adjusted life years on diet, physical activity, and health, 261 traffic-related deaths in, 364 (DALYs) onchocerciasis control program, 217 Global Alliance for Vaccines and attributable to malaria, 206­207 undernourishment measure of, 91 Immunization (GAVI), 156, education food security, 89, 97, 97f, 103, 253 174­175 of girls, 100, 100f, 103 Foundations of Epidemiology Global Burden of Disease (GBD) studies, of medical personnel (Lilienfeld), 228 31, 44, 47­48, 48f, 52 for diabetes, 280 on CVD, 305, 323 for neuroscience, 356­357 Gambia on mental health, 331­333, 332­333t elderly persons cancer in, 290, 296 Global Fund to Fight AIDS, Tuberculosis Alzheimer's disease and, 354 developmental disorders in, 353 and Malaria, 7 causes of death, 56, 56t, 57 diphtheria in, 167 Global Plan to Stop TB 2006­2015, diabetes and, 276, 276f hypertension in, 256 188, 190 falls and, 364 lower respiratory tract infections in, 149, global warming and malaria, 208 emergency obstetric care, 233­234 150­151 Globocan 2000, 52 encephalitis, 353, 355 malaria in, 202, 205 goiter, 95, 341 endomyocardial fibrosis, 312 maternal mortality in, 225 Guinea See also cardiovascular disease (CVD) pregnancy and later risk factors to adult mortality rates in, 37, 38 environmental risk factors children in, 247, 248 cancer in, 290, 296 developmental disabilities and, 139­140 respiratory diseases of children in, 149, meningitis in, 353 diabetes and, 272 150­151, 152­153, 156, 157, 159 tobacco use in, 249 respiratory diseases and, 153 GAVI. See Global Alliance for Vaccines EPI. See Expanded Program and Immunization H. pylori and stomach cancer, 295 on Immunization GBD. See Global Burden of Disease (GBD) Haemophilus influenzae type B (Hib). epilepsy, 341, 353, 355 studies See Hib vaccines research framework, 358 gender differences handicaps. See developmental disabilities treatment of, 351, 357 See also women HBV vaccine, 164, 170­171, 171f, 174 Epstein-Barr virus, 298 alcohol consumption, 335 headaches, 355 eradication of infectious diseases. cancer, 291­292, 291t, 295 health care access issues See specific diseases CVD, 322 See also public health Eritrea diabetes, 272 HIV-infected persons and, 158 female malnutrition in, 229 diarrheal diseases, 115 vaccine-preventable diseases and. vital records in, 45 HIV/AIDS and TB, 184 See vaccines war with Ethiopia, 363 homicide, 363 health care workers. See medical workers esophageal cancer, 299, 300 mental disorders and, 337­338 health systems Ethiopia pertussis, 167 CVD and, 320­323 adult mortality rates in, 37 smoking, 249­250 human resources. See medical workers diabetes in, 272, 273 suicide, 335, 338 maternal mortality and, 231 epilepsy in, 353 traffic-related deaths, 364 surveillance. See surveillance mental disorders in, 332, 333, 334, undernutrition of children, 94, 103 health workers. See medical workers 335, 338 General Health Questionnaire (GHQ), 330 hearing loss, 131, 131t, 135 suicide in, 335 genetics heart disease, 311­312 undernutrition of children in, 94 breast cancer and, 294 See also cardiovascular disease (CVD) war with Eritrea, 363 CVD and, 314, 319 heavily indebted poor countries Expanded Program on Immunization diabetes and, 272 (HIPC), 280 (EPI), 163, 167, 168, 169, 172 hypertension and, 256­257 helicobacter pylori infection eye disease. See blindness; onchocerciasis mental health and, 337 and cancer, 295 Index | 379 hemoglobinopathies, 132, 134 non-Hodgkin's lymphomas and, 298 infant and child mortality, 78­80, hepatitis B, 164, 170­171, 171f, 297 prevalence rates and adult mortality, 80f, 80t See also vaccine-preventable diseases 68­71, 70f, 71t, 74 overall mortality, 77­78 hepatitis C and liver cancer, 296 See also population trends standard age structure, 78, 78f Hib vaccines, 164, 171­172 prevention, 241­242 on patterns of mortality, 81­84 See also vaccine-preventable diseases progression of infection, 241 cluster analysis, 82 for children, 156­157, 159 rates of infection in Africa, 238­239, 244 emerging patterns, 82­84, 83f high blood pressure. See cardiovascular regional variation of, 240­241, 241f as empirical regularities, 84 disease (CVD), subheading: respiratory disease associated with, principal components technique, hypertension 151­152, 154­155, 154­155t 81­82 HIV/AIDS, 237­246 retroviruses and, 239­240 indigenous knowledge. See traditional adult mortality and, 40­41 sexual transmission of, 238, 240 medicine antiretroviral treatment for. See successful prevention programs, 244 infant care. See perinatal conditions antiretroviral drugs (ARV) TB and, 3, 179, 182­185, 183­185f, infant mortality. See child mortality rates blood transfusion and, 3, 241­242 190­191 infectious and communicable diseases breastfeeding and, 242 ARV therapy, 189, 189f See also specific diseases burden of, 1, 3­4, 238f treatment of patients, 187 as causes of death, 49­52 cancer and, 4, 299 undernutrition of children and, 98­100, neurological disorders and, 353­354 See also Kaposi's sarcoma 99f, 101 inherited disorders. See genetics; care and treatment, 158­159, 238f vaccines and microbicides, 242­243 hemoglobinopathies cervical cancer and, 292­293 voluntary counseling and testing, 242 Initiative for Pharmaceutical Technology condom use for prevention, 238, 242 women and, 229 Transfer of South Africa, 321­322 costs and cost-effectiveness of homicide, 362­363 injuries, 361­372 interventions, 244 See also violence See also unintentional injuries; violence CVD associated with, 312­313, 314 South Africa's National Non-Natural burden of, 3, 8­9, 361 dementia and, 354 Mortality Surveillance System, 331, as cause of disability and death, 9, 140 developmental disabilities in children 352, 363, 365 community risk factors, 369 with, 3, 135­136 Hopkins Symptom Checklist-25 data sources on, 362 "development crisis" created by, 4 (HSCL-25), 330 defined, 361 diarrheal diseases and, 108 hospital data collection on maternal drug trafficking and, 369 drug therapy for, 243­244 mortality, 226 family relationships and, 369 epidemiological estimates of deaths HPV vaccine, 293 See also domestic violence caused by, 51­52 human resources. See medical workers peer relationships and, 369 global comparison of prevalence, Human Resources for Health: Overcoming prevention of, 365, 371­372 238­239, 239t the Crisis (Joint Learning risk factors for, 368­370 growing epidemic of, 1, 98, 127, Initiative), 8 societal factors, 369­370 237­238, 244 hunger. See nutrition; undernutrition strategies for control of, 370­371 injecting drug users and, 240 of children insulin, 282, 283­284 interventions for, 189 hyperlipidemia, 317 See also diabetes Kaposi's sarcoma. See Kaposi's sarcoma hypertension, 254­257, 256t, 315­317 intentional injuries. See injuries; violence life expectancy, effect on, 1, 3, 11 See also cardiovascular disease (CVD) International Agency for Research on projections after AIDS, 65­66f, 65­74 hypoglycemia, 278 Cancer (IARC), 291 lower respiratory tract infections and, international attention to health in 151­152, 154­155, 154­155t illicit opiates. See drugs, illegal Sub-Saharan Africa, 6­7 malaria and, 3, 203, 205­206 immunizations. See vaccine-preventable International Classification of Diseases medical workers treating patients and diseases; vaccines; specific diseases and Related Health Problems exposure to, 242 income levels (ICD-10, WHO), 223­224, 228 mental health effects of, 4, 330, 338­339, See also economic growth and International Classification of Functioning, 340t, 341 consequences Disability and Health (WHO), 126, mortality rates and, 1, 53­54, 239 health gains and, 5 126f, 143 children. See child mortality rates violence and injuries linked to, 369­370 International Classification of maternal, 231 INDEPTH Network, 8, 75­86 Impairments, Disabilities and UNAIDS/WHO model on mortality on comparative levels of mortality, Handicaps (WHO), 126, 126f estimates, 51­52 77­81, 77t International Conference on Population mother-to-child transmission, 135, 242 adult mortality, 80­81, 81t and Development (ICPD), 223 neurological impact of, 353­354 crude death rates, 78, 79f, 79t International Diabetes Federation, 273 380 | Index International Union against Tuberculosis undernutrition of children in, 101 agricultural practices and, 199 and Lung Disease (IUATLD), 185 war in, 363 anemia and, 196, 203, 205, 206 interventions. See drugs; lifestyle; specific life expectancy, trends in, 11, 277 artemisinin to treat, 208 diseases and conditions See also mortality rates cerebral malaria, 136, 200, 203, 354 intrauterine growth retardation, 134 comparison of Sub-Saharan Africa with children and, 195, 202 iodine deficiency, 88, 92t, 95, 138 other regions, 12, 12t, 13, 13t, 14f climate determinants of, 153, 197, 198f iron deficiency, 92t, 138 estimates of, 33 combined indirect effects on all-cause See also anemia HIV/AIDS and, 1, 3, 11 mortality, 206, 207­208f iron overload and liver cancer, 297 projections after AIDS, 65­66f, 65­74 demographic surveillance systems (DSS) isoniazid preventive treatment (IPT), 188 of women after AIDS, 66­68, 67­68f to monitor, 200, 206, 208 ivermectin, 218­220, 219f lifestyle, 2 diagnosis of, 200 See also specific lifestyle choice (e.g., alcohol distribution in Africa, 197­199 Joint National Committee on Detection, consumption, tobacco use) drug reactions to commonly used Evaluation, and Treatment of High chronic diseases and, 247­262 antimalerial drugs, 203 Blood Pressure VI (JNC VI), 315, See also risk factors for drug resistance and, 199, 208 317, 323 noncommunicable diseases economic loss attributable to, 206­207 Joint United Nations Programme on physical exercise and, 251­252 epidemiological estimates of deaths HIV/AIDS. See United Nations policy initiatives to promote healthy caused by, 50, 50t Programme on HIV/AIDS lifestyle, 261­262 HIV/AIDS and, 203, 205­206 (UNAIDS) urbanization. See urbanization low birthweight and, 204­205, 206 life tables macroeconomic impact of, 206­207 Kaposi's sarcoma, 4, 294­295, 300 causes of death, construction of, 48­49, malnutrition and, 196, 204, 205 Kenya 49f, 61 Mapping Malaria in Africa Project, adult mortality rates in, 35, 35t model life tables, use of, 76, 85 197­199 cardiovascular disease in, 308, 312 WHO mortality and life table estimates, mortality due to, 195, 200­206 child mortality rates in, 15, 15f 33, 40, 307 child deaths. See child mortality rates diabetes in, 280 liver cancer, 296­297 consequential mortality, 203­204 esophageal cancer in, 299 low birthweight indirect mortality, 204­206 HIV/AIDS prevalence estimates and blood pressure problems associated measuring malaria-specific mortality, demographic impact in, 70, 71, 73 with, 248 200­201, 201­202t hypertension in, 256, 315 effect of averting, 102­103 rates, 209 malaria in, 195, 196f, 198, 199, 201, malaria and, 204­205, 206 trends in, 207­208 202, 205 smoking during pregnancy and, 247­248 neurological disability and, 203­204 maternal mortality levels in, 226, 234 undernutrition of children and, 90, 91f populations at risk, 198­199, 199t mental health studies in, 330 lower respiratory tract infections (LRTI) poverty and, 206­207 respiratory tract infections in, 1, burden of, 150 pregnancy and, 196, 204­205 149, 150 cost of interventions, 158 public health effects of, 196, 196f suicide in, 335 diagnosis of, 153­154, 154t refugee populations and, 199 TB in, 189f epidemiology of, 149­151 severe malaria anemia (SMA), 200, 203 tobacco use in, 249 etiology of, 152­155 transfusions and, 203 traffic-related deaths in, 364 Hib conjugate vaccine and, 156­157, 159 undernutrition and, 96, 196, 204, 205 vital registration system in, 57 HIV/AIDS and, 151­152, 154­155, Malaria Atlas Project, 209 kidney disease and diabetes, 278 154­155t Malawi incidence of, 150 adult mortality rates in, 35, 35t, 38 landlocked countries and undernutrition interventions, 155­159 cancer in, 290 of children, 93, 93f mortality due to, 51, 149 diabetes in, 280 landmines, 363 prevention of, 153 HIV/AIDS in, 159, 206 lead exposure, 140 WHO management strategy, TB in conjunction with, 184, 190 learning disabilities, 132, 353 155­156, 159 Kaposi's sarcoma in, 294 See also developmental disabilities lung cancer, 299­300 malaria in, 206 Lesotho medical personnel in, 8 alcohol consumption in, 335 macroeconomic impact of malaria, respiratory diseases of children in, women's life expectancy, after AIDS in, 68f 206­207 155­156 Liberia malaria, 195­213 tobacco use in, 248 conflict in, 8 adverse drug reactions to commonly used undernutrition of children in, 94 infant mortality rates in, 13 antimalerial drugs, 203 males. See gender differences Index | 381 Mali maternal undernutrition, 87, 88t research needed on, 343­344 adult mortality rates in, 38 Mauritius schizophrenia, 332­333 cancer in, 290 infant and children under age 5 mortality genetic vulnerability and, 337 malaria in, 199 rates in, 13, 109 specific data on Sub-Saharan Africa, 334 malnutrition. See undernutrition of children maternal mortality levels in, 228 treatment of, 342 Mapping Malaria in Africa Project, 197­199 MDGs. See Millennium Development Goals scope of, 329­330 maternal conditions measles, 137 suicide. See suicide See also abortion See also vaccine-preventable diseases treatment of, 342­343 diarrheal diseases of children and quality burden of, 137, 165­166 wars and conflicts as cause of, 9 of care, 114 vaccine for, 166, 166f mental retardation, 130 HIV/AIDS and, 3­4 Mectizan. See ivermectin See also developmental disabilities See also breastfeeding medical workers causes of, 135 malaria and, 204 See also physicians Down syndrome, 134 MDGs on, 234 need for, 8, 98 iodine deficiency and, 138 mental disorders of mother, effect on numbers and types of, 8, 98, 99f midwives. See childbirth children, 339 training. See training of medical personnel migraine headaches, 355 undernutrition of children and quality vacancy rates, 8, 8t Millennium Development Goals of care, 91­92, 97, 100, 100f meningitis, 136­137, 165, 172, 353, 355 (MDGs), 7b undernutrition of mother, 87, 88t mental disorders, 329­346, 345­346t on hunger and undernutrition, maternal mortality, 223­236 See also specific disorders 104­105, 105f abortion and, 229 anxiety disorders, 334 importance to all countries, 7, 7b access to skilled attendance and, 231, bipolar disorders, 332 on infant and children under age 5 231f, 233 burden of, 3 mortality, 15, 117, 234 biological-demographic variables for, 230 childhood, 335 on maternal health, 234 causes of, 228­229, 229f treatment of, 343 poverty reduction and, 4­5 costs of childbirth, 230 common mental disorders, 334 on vaccine-preventable diseases, 175 data collection on, 32, 223­226 comorbidity, 339­341 mortality rates, 11­14, 31­42 direct sibling-history method, 224, 225 consequences of, 339­341 See also specific diseases and conditions health services data, 226 data sources on, 330­331 AIDS-related. See population trends household surveys, 224, 225­226 depression, 331­332 causes of death, 43­58, 290t measures of, 224 identification of, 331 See also causes of death population-based data, 224, 225 life events and, 337 comparison of Sub-Saharan Africa with sisterhood method, 224, 225 physical problems resulting from, 339 other regions, 12, 12t sources of data, 224­226 postpartum depression, 338 crude mortality rates, 78, 79f, 79t vital registration, 224, 225 specific data on Sub-Saharan Africa, 334 age-standardized mortality vs., 307, 308f definitions, 223­224 substance abuse and, 339, 341 data sources on, 32­33 determinants of, 229­232 treatment of, 342 INDEPTH Network on comparative levels emergency care and, 233­234 women, 337­338 and patterns, 75­86 epidemiological estimates of deaths epidemiology of, 331­341 See also INDEPTH Network caused by, 51 etiology of, 336­339 indicators of levels and trends, 11­12 health systems and, 231 GBD 2000 study on, 331­333, 332­333t levels and trends, 12­14, 35­39, HIV/AIDS and, 231 gender differences and, 337­338 35t, 36­39f household and community genetic vulnerability and, 337 estimates of levels, 33­35, 34t, 35f characteristics, 230 HIV/AIDS and, 4, 330, 338­339, 340t, 341 maternal, 223­236 interventions, 232­234 interventions, 341­343 See also maternal mortality late maternal death, 224 malnutrition and, 336­337 subregional differences in, 12­13 levels, trends, and differentials, 54, obsessive-compulsive disorders, 333 surveillance and, 8 226­227t, 226­228 panic disorders, 332 mosquitoes. See malaria levels of delay, model of, 233­234, 233b policy and service implications, 344 motor disabilities, 131 lifetime risk of maternal death, 224 posttraumatic stress disorder (PTSD), 333 See also developmental disabilities malnutrition-infection syndrome, specific data on Sub-Saharan Africa, motor vehicle injuries. See traffic-related 230­231 334­335 injuries national policies and investments, effect treatment of, 342­343 Mozambique of, 231­232 poverty and, 337, 339 landmines in, 363 "near-miss population" and, 234 primary prevention of, 341­342 malaria in, 199, 205 risk factors for, 230 promotion of self and, 341 physical activity of children in, 252 382 | Index safety program in, 371 trauma and, 355 diabetes and, 253, 277 undernutrition of children in, 94 treatment of, 357­358 prevalence of, 319, 319t Multiple Indicator Cluster Surveys newborns. See child mortality rates; as risk factor for chronic disease, 255 (MICS), 17 perinatal conditions obstetric care. See childbirth; maternal myocardial infarction, 310­311 disorders of. See congenital disorders conditions See also cardiovascular disease (CVD) New Partnership for Africa's Development onchocerciasis, 215­222 (NEPAD), 104, 322 control programs, 217­220 Namibia Nigeria disease burden of, 215­216 adult mortality rates in, 37 adult mortality rates in, 37 ivermectin treatment, 218­220, 219f malaria in, 201 alcohol consumption in, 335 partnership of control programs, 217 safe water and sanitation in, 98 cardiovascular disease in, 307 population at risk and population tobacco use in, 249 clustering of risk factors in, 260, 260t infected, 216­217 women's life expectancy after AIDS, 66, 67f developmental disorders in, 353 Rapid Epidemiological Mapping of, 216 NDGO Coordination Group for Ivermectin diabetes in, 272, 273, 276, 280 socioeconomic burden of, 216 Distribution, 219 Down syndrome in, 134 transmission of, 215 neonatal conditions. See childbirth; child HIV/AIDS in, 240 vector control, 217­218, 218f, 220t mortality rates; perinatal conditions hypertension in, 254­255, 257, 315, 316 Onchocerciasis Control Program (OCP), neonatal tetanus, 168 life expectancy projections, after AIDS, 65f 215, 217, 218, 219, 220, 221 NEPAD. See New Partnership for Africa's liver cancer in, 296 Onchocerciasis Elimination Program for Development lower respiratory tract infections in, 149 the Americas (OEPA), 219 neural tube defects, 134 malaria in, 205 oral rehydration therapy (ORT), 107 neurological disorders, 351­359 migraine headaches in, 355 Alzheimer's disease and dementias, 354 neural tube defects in, 134 panic disorders, 332 burden of, 3 neurological disorders in, 355 Parkinson's disease, 354­355 in children with HIV/AIDS, 4 population projections after AIDS, 62, 62f pericarditis, 314 consequences of, 356 respiratory diseases of children in, 156 perinatal conditions cost-effectiveness of interventions, 358 strokes in, 310 breastfeeding. See breastfeeding data sources on, 352 tobacco use in, 248 developmental disabilities and, developmental. See developmental traditional medicine in, 356 134, 141 disabilities undernutrition of children in, 94 nutritional care, 91­92 diagnostic facilities and equipment, need urban lifestyle and lack of physical tetanus and, 137, 168 for, 357 activity in, 252 pertussis epidemiology of, 352­355 noncommunicable diseases burden of, 167­168 epilepsy. See epilepsy See also specific diseases and conditions vaccine for, 163, 168­169 etiology and determinants of, 355 burden of, 2, 316 See also vaccines headaches, 355 increase in, 2, 267 pharmaceuticals. See drugs HIV/AIDS and, 353­354 risk factors for, 246­262 phenobarbital, 357 interventions, 351, 357­358 See also risk factors for physical activity, 251­252 malaria and, 203­204 noncommunicable diseases CVD and, 317 movement disorders, 354­355 non-Hodgkin's lymphomas, 298­299 diabetes and, 277 need for medical worker specialists nutrition physicians on, 356­357 cholesterol and, 260 See also medical workers neurodevelopmental disabilities, 134 CVD and, 314 certification of cause of death by, 57 Parkinson's disease, 354­355 hypertension and, 255 shortage of, impact on treatment, policy issues, 358­359 improvements from agricultural 356­357 prevention of, 355, 357 productivity, 103 pneumonia, 151­152, 167 research framework, 358 lifestyle and, 250­251, 251f See also lower respiratory tract infections resource issues, 356­357 policy making on, 103, 104­105 (LRTI); respiratory diseases of scope of, 351­352 pregnant women, 247 children sexually transmitted infections and, 354 sodium levels, 255, 296, 312 policy making stroke typical diet in Sub-Saharan Africa, abortions and, 231­232, 232f See also stroke 250­251, 255­256 initiatives to promote healthy lifestyle, burden of disease, 354 261­262 interventions, 357 obesity, 253, 254t mental disorders and, 344 research framework, 358 breast cancer and, 294 neurological disorders and, 358­359 TB and, 354 CVD and, 315, 319 nutrition and, 103, 104­105 Index | 383 polio psychiatric disorders. See mental disorders dyslipidemia and, 258­260, 259t burden of, 164­165 PTSD. See posttraumatic stress disorder health services' requirements for global elimination, goal of, 165, 173 public health management of, 260­261, 260t vaccine for, 163, 165 malaria, effect of, 196, 196f injuries and, 368­370 See also vaccine-preventable diseases prenatal and infant care, effect of, 91­92, obesity and, 253, 254t population trends, 5­6, 6f 102­103 physical activity, lack of, 251­252 projections after AIDS, 59­74 surveillance. See surveillance promotion of healthy lifestyle to counter, birth rate projections, 63­64, 64f 261­262 data sources, 59­61 radiation therapy for cancer, 290f smoking and, 248­250, 249­250t effects on, decomposed, 61­65, 62f, 63t RAMOS (Reproductive Age Mortality river blindness. See onchocerciasis fertility projections, 62­63 Studies), 225 road traffic injuries. See traffic-related HIV prevalence and, 68­71, 70f, 71t, 74 rape, 9, 338 injuries life expectancy projections, 65­70f, Rapid Epidemiological Mapping of Roll Back Malaria movement, 208 65­74 Onchocerciasis, 216 Romania and abortion, 231­232, 232f sex and age ratios for life expectancy, refugee populations, 8­9 rotavirus, 113­114, 113­114t 67­68, 68f diarrheal diseases and, 115 See also diarrheal diseases postpartum depression, 338 malaria and, 199 immunization for, 172 posttraumatic stress disorder (PTSD), 333 regional comparisons rubella specific data on Sub-Saharan Africa, adult mortality, 35­39, 38­39f See also vaccine-preventable diseases 334­335 child mortality under age of five, 18­21, congenital rubella, 135 treatment of, 342­343 18­21f vaccination, 135 poverty leading causes of death, 54, 54t Rwanda CVD and, 305 registration of births, deaths, marriages, adult mortality rates in, 37 developmental disabilities and, 126, etc. See vital events registration alcohol consumption in, 335 138­139 religious and cultural customs cancer in, 295, 296 diarrheal diseases and, 114 maternal mortality and, 230 conflict in, 8, 9 injuries and, 369­370 neurological health and, 351, 358 depression in, 334 lower respiratory tract infections and, 151 violence and, 370 diarrheal diseases in refugees from, 115 malaria and, 206­207 Reproductive Age Mortality Studies infant mortality rates in, 13 maternal mortality and, 230 (RAMOS), 225 malaria in, 206 mental disorders and, 337, 339 research and development PTSD in, 335 reduction, 4­5 mental disorders, 343­344 war in, 363 undernutrition and, 87­88 vaccines, 173 violence and, 369­370 respiratory diseases of children Safe Motherhood Initiative (1987), 228 Poverty Reduction Strategy approach, 230 as causes of death, 51 safe water, access to, 98, 98f, 114 pregnancy cost of interventions, 158 sanitation improvements See also childbirth; maternal mortality; Hib conjugate vaccine and, 156­157, 159 diarrheal diseases and, 114­115, 115f, 115t prenatal care programs lower respiratory tract infections, 3, undernutrition of children and, 98, 98f developmental disabilities and, 134 149­162 schizophrenia. See mental disorders malaria and, 196, 204­205 See also lower respiratory tract Senegal nutrition during, 247 infections (LRTI) diabetes in, 272 risk factors for chronic diseases from, pneumonia, 151­152, 153 liver cancer in, 296 247­248 S. pneumoniae conjugate vaccine and, malaria in, 202 smoking during, 247­248 157­158 pertussis in, 167, 168 vaccines during, 169, 169f retroviruses. See HIV/AIDS physical activity of adolescent girls in, 252 premature births. See childbirth; low rheumatic heart disease (RHD), 311 severe malaria anemia (SMA), 200, 203 birthweight See also cardiovascular disease (CVD) sexually transmitted infections prenatal care programs risk factors for noncommunicable diseases, HIV/AIDS. See HIV/AIDS See also maternal conditions 2, 247­262 neurological impact of, 354 access to, 98, 99f See also specific risks (e.g., cholesterol levels, sexual violence and abuse, 338 nutritional care, 91­92, 102­103 obesity) See also rape prescriptions. See drugs antenatal influences on, 247­248 shigella. See diarrheal diseases prevention. See specific diseases clustering of multiple risks, effect of, sibling survival as source of mortality and conditions 260­261, 260t data, 32 prostate cancer, 297­298, 300 CVD and, 314­315t, 314­319 sickle cell anemia and related disorders, protease inhibitors, 243 diabetes and, 273­277 134, 313, 355 384 | Index Sierra Leone tobacco use in, 248, 249­250 diabetes in, 272, 273, 276, 279­280, 322 infant mortality rates in, 13 vital registration system in, 57, 109, diet in, 251, 256 malaria in, 202 290, 300 HIV/AIDS in, 240 undernutrition of children in, 101 Special Programme for Research and hypertension in, 255 SMA. See severe malaria anemia Training in Tropical Diseases malaria in, 201, 202 smallpox eradication, 163 (WHO), 216, 219 maternal mortality levels in, 226, 234 smoking. See tobacco use S. pneumoniae conjugate vaccine, 157­158 mental disorders in, 341 socioeconomic context, 4­5, 4f stigma and neurological disorders, 351, 358 noncommunicable disease burden in, 252, CVD, increase in, 321 stillbirths. See child mortality 316, 317f onchocerciasis, burden of, 216 stomach cancer, 295­296 physical activity in, 317 sodium in diet, 255, 296, 312 Stop TB Strategy, 184, 185, 186b, 187, 188 stomach cancer in, 295 Somalia and undernutrition stroke, 309­310, 310t, 314, 315, 323, 354 suicide in, 335 of children, 101 See also cardiovascular disease (CVD); TB in, 182 South Africa neurological disorders traffic-related deaths in, 364 adult mortality rates in, 36­37, as cause of death, 306 undernutrition of children in, 94 36­37f, 38 hypertension and, 355 unintentional injuries in, 364 Afrikaans-speaking people and high interventions, 357 T-cell lymphomas, 298 cholesterol levels in, 260 stunting tetanus breast cancer in, 294 See also undernutrition of children burden of, 137, 168 cardiovascular disease in, 309, 310, 313 in children under age five, 93, 93f elimination goal, 168 low birthweight and, 248 economic loss due to, 88 maternal and neonatal, 169, 169f data collection in, 32, 109 gender differences, 94 vaccine for, 163, 168­169 depression in, 334 regional patterns in, 93 TMP-SMX prophylaxis, 158­159 developmental disabilities studies in, 128 substance abuse. See alcohol consumption; tobacco use, 248­250, 318t diabetes in, 252, 272, 273, 278­279, 280 drugs, illegal adolescents and, 248, 249t ethnic differences and, 273, 276 Sudan adults and, 249­250, 250t diagnostic facilities and equipment in, 357 alcohol consumption in, 335 cigarette smoking per adult, 300, 300f Down syndrome in, 134 diabetes in, 272 CVD and, 317 esophageal cancer in, 299 meningitis in, 353 esophageal cancer and, 299 fetal alcohol syndrome in, 139 refugee populations of, 9 marketing, effect of, 248­249, 250 "founder effect" in, 260 suicide, 363 pregnant women's smoking, effect of, HIV/AIDS treatment in, 158 data specific to Sub-Saharan Africa, 247­248 HIV children with developmental 335, 336t price increases and, 250 disabilities in, 135 gender differences, 338 WHO interventions, 261 homicide rate in, 365 HIV positive and, 3 Togo and adult mortality rates, 38 hypertension in, 256, 257, 316 South Africa's National Non-Natural traditional medicine injuries in, 364­368, 370 Mortality Surveillance System, CVD, treatment by, 322 Lead Programme on Crime, Violence 331, 352 neurological disorders, treatment by, 356 and Injury, 371­372 sulfonamides, 203 traffic-related injuries, 3, 364 learning disabilities in, 132 surveillance, 8, 33, 307 prevention of, 370 lung cancer in, 300 surveys as method of collecting data on training of medical personnel malaria in, 199, 201, 202 child mortality rates, 16­17 for diabetes, 280 mortality registration in, 8 Swaziland for neuroscience, 356­357 National Non-Natural Mortality Kaposi's sarcoma in, 294 trauma and neurological disorders, 355 Surveillance System, 331, 352, life expectancy projections, after AIDS, tuberculosis (TB), 179­193 363, 365 65f, 66, 70 burden of, 169­170, 179 neural tube defects in, 134 population projections, after AIDS, 62, 62f children and, 182 nutrition in, 251, 251f Sydenham's chorea, 354 complementary strategies, 188 obesity in, 253 syphilis constraints on programs in physical activity in, 252, 277, 317 congenital syphilis, 135 controlling, 190b prostate cancer in, 297 neurological impact of, 354 cotrimoxazole treatment, 190 respiratory diseases of children in, 153, DALYs and, 179 155, 157 Tanzania diagnosis of, 181­183 stomach cancer in, 295 Adult Morbidity and Mortality Project DOTS strategy for treatment, 184, stroke in, 354, 355 (AMMP) in, 33, 45, 316 185­190, 187f, 188 suicide in, 335 cholesterol levels in, 258­260 drug resistance and, 185 Index | 385 tuberculosis (Continued ) interventions needed to eliminate, urbanization epidemiology of, 183­190 101­103, 102t CVD and, 315, 320 extrapulmonary, 182 policies and resources needed for, diabetes and, 276­277 HIV/AIDS and, 3, 179, 182­185, 104­105 hypertension and, 255, 315­316, 316t 183­185f, 190­191 scale and commitment of, 103­104 lifestyle and, 253, 255 ARV therapy, 189, 189f life cycle and cross-generational patterns U.S. Census Bureau, 32 treatment of patients, 187 of, 90­91, 91f birth rate projections of, after AIDS, incidence rates, 170, 179, 184t, 190 malaria and, 96, 196, 204, 205 64, 64f isoniazid preventive treatment (IPT), 188 mental disorders and, 336­337 population projections of, after AIDS, lower respiratory tract infections and, 155 outcome indicators of, 90­91, 90t, 59­74 microbiology of, 180 91f, 92t See also population trends mortality due to, 170, 184 policy making on, 103­104 neurological impact of, 354 quality of care and, 91­92, 97, 100, 100f vaccine-preventable diseases, 163­177 pathogenesis of, 181 respiratory diseases of children and, 153 See also vaccines; specific diseases and preventive treatment, 188 rural vs. urban areas, 94, 95f conditions pulmonary, 181­182 status and trends in, 93­101 burden of, 164­172 Stop TB Strategy, 184, 185, 186b, 187, 188 under age 5 mortality, 95­96, 96f DALYs lost from, 164, 164t transmission of, 180­181 daily dietary supply available and food epidemiological estimates of deaths vaccination, 163 production index, 97­101, 97f caused by, 50­51 Tunisia regional distribution, 93­94, 93­94f Expanded Program on Immunization. cerebrovascular disease in, 309 wealth of household, effect of, 94­95, 95f See Expanded Program on risk factors for CVD in, 315 UNICEF. See United Nations Children's Immunization (EPI) Fund inequities in access to vaccines Uganda unintentional injuries, 361, 363­364 (vaccine gap), 173­174 adult mortality rates in, 37 See also burn-related injuries; drowning; mortality due to, 164t cancer in, 290, 295, 299 fall-related injuries; traffic-related research and development agenda, 173 cardiovascular disease in, 312 injuries vaccines depression in, 334 United Kingdom Prospective Diabetes See also vaccine-preventable diseases; HIV/AIDS in, 8, 205 Study, 283 specific diseases population projections and, 61 United Nations BCG (bacillus Calmette-Guérin), hypertension in, 315 See also Millennium Development Goals 170, 189 injuries in, 364­368, 366­368t (MDGs); specific UN agencies childhood diseases, 173, 173f Kaposi's sarcoma in, 294 life expectancy projections, 3 cost-effectiveness of programs, 135, 163, malaria in, 205 United Nations Children's Fund (UNICEF) 166, 169 meningitis in, 353 Alma Ata Declaration, 228 decrease in rates, 9 non-Hodgkin's lymphoma in, 298 on maternal mortality and morbidity, diphtheria-tetanus-pertussis (DTP), UK Clinical Interview Schedule, 330 224, 229, 234 166­169, 169f, 172, 173 UNAIDS. See United Nations Programme Multiple Indicator Cluster Surveys financing of, 174­175 on HIV/AIDS (MICS), 12, 108 future challenges for, 172 undernutrition of children, 87­106 on nutritional status of children, 88, HBV, 164, 170­171, 171f, 174 See also stunting 89f, 103 Hib conjugate vaccine and, 156­157, 159 breastfeeding and, 92, 93, 100, 101f on vaccine-preventable diseases, 173 history in Africa of, 163­164, 164f burden of disease attributable to, 87, United Nations Development Programme inequities in access to (vaccine gap), 88t, 91 (UNDP) and onchocerciasis 173­174 causes of, 88­89, 89f control, 216, 217, 219 problems and barriers to adoption, 174 cognitive function and, 138, 139 United Nations Population Division research and development agenda, 173 data sources on, 92 See also population trends rubella immunization of adolescent girls developmental disabilities and, 137­139 on mortality rates, 12, 33, 40 or women of reproductive age, 135 diarrheal diseases and, 96 population projections of, 199 S. pneumoniae conjugate vaccine, 157­158 economic effects of, 87­88, 95, 100­101, after AIDS, 59, 60, 73 successful programs of, 137 102, 102t, 105f United Nations Population Fund on tetanus immunization during pregnancy, food security and, 89, 97, 97f, 103 maternal mortality, 224, 234 137, 168 gender and age of children, 94 United Nations Programme on HIV/AIDS unsafe administration of, 173 "hidden hunger," 91, 95 (UNAIDS), 51, 238 vector control and onchocerciasis, indicators of, 90­92, 90t population projections, after AIDS, 60, 217­218, 218f, 220t input indicators of, 91­92 69­71, 73 verbal autopsies, 57, 76, 200, 307 386 | Index violence, 361­372 tobacco use of, 249­250 on tobacco control, 261 See also injuries during pregnancy, 247­248 on undernutrition of children, 92 against women, 338, 370 violence against, 338, 370 on vaccines, 173 See also rape World Bank See also Expanded Program on burden of, 3 birth rate projections of, after AIDS, Immunization (EPI) as cause of death, 362 64, 64f on violence-related deaths, 362 culture and, 370 CVD control and, 323 on vital records, 45 family relationships and. See domestic development assistance for health needs, 7 on war between Ethiopia and Eritrea, 363 violence onchocerciasis control and, 216, 217, 219 World Health Report 2002: Reducing Risks, homicide, 362­363 population projections of, 6 Promoting Healthy Life, 336 prevention of, 370­371 See also population trends World Population Prospects (UN), 33 viral encephalitis, 353, 355 after AIDS, 59, 61­74 World Report on Violence and Health viruses as leading cause of lower World Conference on Women (Fourth), 223 (WHO), 371 respiratory tract infections, World Development Report 1993: Investing 152­153 in Health (World Bank), 163 yellow fever, 163, 170, 170f vision loss. See blindness World Fertility Survey (WFS), 12, 76 See also vaccine-preventable diseases vital events registration World Health Organization (WHO) young adults. See adolescents and young adult mortality rates and, 32, 39 See also Global Burden of Disease (GBD) adults causes of death and, 45­47, 46t, 57, 109 studies; names of programs starting Youth Tobacco Survey, 248, 249t child mortality rates and, 16 with "WHO" lack of data, effect of, 8, 32, 39­40, 57, on alcohol consumption, 335 Zambia 109, 290 Alma Ata Declaration, 228 adult mortality rates in, 38 maternal mortality rates and, 224, 225 on antiretroviral drugs, 322 cardiovascular disease in, 308 vitamin A deficiency, 92t, 95, 138 burden of disease estimates of, 267 HIV/AIDS in on cancer, 298 life expectancy projections and, wars and conflicts on chronic disease management, 261 71­72, 72f burden of injuries from, 8­9.9t, 363 on coordinated patient care, 261 mental disorders and, 339 child soldiers in, 140 on CVD, 323 prevalence rates, 70 displaced persons from, 8­9 on diabetes, 267 life expectancy in, 13 epidemiological estimates of deaths on diet, physical activity, and health, 261 medical personnel in, 8 caused by, 52 on epilepsy, 353, 357 mental disorders in, 339, 341 weight. See low birthweight; obesity Expanded Program on Immunization. stroke in, 354 Wellcome Trust, 209 See Expanded Program on vital records in, 45 WHO. See World Health Organization Immunization (EPI) Zimbabwe WHO Global Strategy for on injuries, 361, 368, 369, 371 adult mortality rates in, 36, 36f, 38 Noncommunicable Diseases, 321 Injuries and Violence Prevention cancer in, 290, 297, 299 whooping cough. See pertussis Department, 371 depression in, 334 WHO­World Bank High-Level Forum on International Classification of Functioning, diabetes in, 272, 279, 280 the Health MDGs, 7 Disability and Health, 126, 126f, 143 headaches in, 355 women International Classification of HIV/AIDS in, 239 See also gender differences Impairments, Disabilities and hypertension in, 255 anemia and, 95, 229 Handicaps, 126, 126f Kaposi's sarcoma in, 294 cancer and, 292­294 on maternal mortality, 224, 234 learning disabilities in, 132 death in childbirth. See maternal on mortality and life table estimates, 33, life expectancy in, 13, 68f, 69 mortality 40, 307 malaria in, 201 education of girls, 100, 100f, 103 onchocerciasis control program, 217 malnutrition in, 88 HIV/AIDS and, 229 on polio eradication, 165 neurological disorders in, 352 TB in conjunction with, 184 on poverty, 369 non-Hodgkin's lymphoma in, 298 life expectancy of, after AIDS, 66­68, on respiratory diseases of children, population projections after AIDS, 63 67­68f 155­156, 159 strokes in, 310 maternal mortality, 223­236 on safe water and sanitation, 114 tobacco use in, 248 See also maternal mortality on smoking, 300 vital registration system in, 57, 109 obesity rates in, 319, 319t Stop TB Strategy, 184, 185, 186b, 187, 188 women pregnancy care. See prenatal care survey for global burden of disease, 32 life expectancy of, after AIDS, 66, 67f programs on TB, 185 postpartum depression of, 338 rape. See rape on tetanus, 168 zinc deficiency, 92t Index | 387 T his new edition of Disease and Mortality in Sub-Saharan Africa represents a dramatic advance- ment in knowledge since the publication in 1991 of the first edition. The potential impact of HIV/AIDS was anticipated in that year, but the current volume documents the depth and breadth of the burden that epidemic is inflicting on Africa. Updates to the data on morbidity and mortal- ity from 1990 to 2004 are put into context with discussions also of the conflicts and wars, the famines and droughts, the migration of human resources, and the global economic downturn that have affected the health of the population in the region. The goal of this second edition remains the same as that of the first: to provide policy makers with a comprehensive resource of high-quality information on the nature and extent of health problems in Sub- Saharan Africa and the health care strategies that show the most promise. Good data do not ensure good decisions, well-designed programs, or effective implementation, but they are a necessary first step. This volume reflects the growth in technical capacity for data collection and analysis of health statistics since 1991. Many new sources of health and demographic information have become available as a result of unprecedented international interest in health conditions in Sub-Saharan Africa. The Millennium Development Goals, described in the Millennium Declaration signed by 189 countries in 2000, focused the world's attention on specific targets--including health outcomes--and deadlines to reduce global poverty. The year 2015, the deadline for achieving the three major health goals, is just around the corner. Sub-Saharan Africa is not on track to reach any of the public health Millennium Development Goals. HIV/AIDS has taken its toll, leading to a stalling or a reversal of the gains made in health indicators, including infant and under-five mortality rates and life expectancy at birth. The epidemic has not only increased the spread of communicable disorders--such as tuberculosis, malaria, respiratory tract infections, and diarrhea--but has also affected the epidemiology of cancers, such as Karposi's sarco- ma, mental health disorders, such as depression, and neurological disorders, such as dementia. Not all trends have been negative. Uganda has seen a decrease in the prevalence of HIV/AIDS within its borders, southern Africa has virtually eliminated measles mortality, and across Africa onchocerciasis control improved greatly during the 1990s. But the continued success of such efforts depends on both the monitoring of disease indicators and the effectiveness of programs to address them. The year 2015 looms and presents a large challenge, so high priority must be given to the continuing improvement of disease surveillance, the timely sharing of public health information, and wise policy decisions. This book will be an invaluable resource to those working in the health sector--in epidemiology, pub- lic health, population and reproductive health, and health policy--whether in government, research, academia, consulting, or operations. ISBN 0-8213-6397-2